[00:00.000 --> 00:00.640]  ...place to do it.
[00:01.360 --> 00:11.240]  And also, simultaneously, next door, in the Lamar Room at midnight, is a screening of Dark Web, directed by Alex Winter about the Silk Road case.
[00:12.060 --> 00:15.880]  But right now, I'm going to introduce Elissa Shevinsky, a.k.a. Lady Boss, who is going to be doing a talk about censorship, social media, and the presidential election.
[00:22.540 --> 00:23.380]  Very topical.
[00:23.700 --> 00:25.960]  So, take it away, Lady Boss.
[00:26.180 --> 00:26.520]  Thank you.
[00:26.760 --> 00:27.840]  Thank you so much.
[00:30.000 --> 00:32.040]  And thank you all for coming.
[00:32.620 --> 00:35.180]  I'm really glad that we're discussing this.
[00:35.320 --> 00:41.060]  This is such an important topic, increasingly so, with each election.
[00:41.180 --> 00:49.460]  So, I'm really glad for us to talk about it here at HOPE, where we can be, how do I say, we can be very honest.
[00:49.940 --> 00:57.380]  I posted a tweet that I might destroy, like, all my future likely job prospects in this one talk.
[00:57.380 --> 01:07.240]  So, I'm glad that you're here to be here with me for this, as I complain about pretty much all the major political and technological organizations.
[01:07.240 --> 01:11.300]  Any bias in this talk is brought to you by me.
[01:11.480 --> 01:15.580]  That's my hashtag and my handle on Twitter.
[01:15.980 --> 01:17.060]  Some disclosures.
[01:17.060 --> 01:20.480]  I was formerly on Rand Paul's Tech Advisory Council.
[01:20.480 --> 01:27.220]  I'm talking to folks who, yeah, I'm talking to folks who worked for both Clinton and Trump.
[01:27.540 --> 01:36.400]  I'm very strongly pro-Clinton, but just you would not be able to tell this from this talk, because I'm pretty upset about the bias in her favor.
[01:37.880 --> 01:41.220]  We accept political endorsements by the press.
[01:41.560 --> 01:42.780]  We're used to that.
[01:42.960 --> 01:43.900]  It's fine.
[01:44.400 --> 01:47.020]  Or if it's not fine, we're certainly used to it.
[01:47.160 --> 01:51.860]  And one thing that is really nice is they disclose their bias.
[01:52.240 --> 01:55.720]  The Post officially endorses Donald Trump, they declare.
[01:55.720 --> 02:00.980]  And the Daily News explains why they choose Hillary Hill over Bernie.
[02:02.860 --> 02:05.460]  And we see op-eds that reflect this.
[02:05.780 --> 02:08.860]  There's no pretense of neutrality.
[02:09.180 --> 02:13.520]  There's the understanding that there's editorship and curation.
[02:14.640 --> 02:17.780]  But social media is supposed to be neutral.
[02:18.120 --> 02:24.120]  It's this beautiful fantasy that we've created our Facebook pages.
[02:25.000 --> 02:28.240]  And indeed, the companies support this.
[02:28.540 --> 02:34.200]  Facebook has officially denied that it would ever try to influence the election.
[02:34.660 --> 02:37.860]  They haven't denied that they ever would influence the election.
[02:38.200 --> 02:41.440]  They've denied that they would try, not that they would succeed.
[02:44.690 --> 02:48.030]  So tell us more about this neutrality, Mark Zuckerberg.
[02:48.250 --> 02:49.630]  Well, we have guidelines.
[02:50.310 --> 02:52.210]  They have guidelines, so it's fine.
[02:52.450 --> 03:01.910]  We have rigorous guidelines that do not permit the prioritization of one viewpoint over another or the suppression of political perspectives.
[03:03.470 --> 03:04.590]  Tell me more.
[03:04.890 --> 03:06.910]  Well, we have the trending section.
[03:08.070 --> 03:12.410]  As you can see, (((Mark Cuban))) is criticizing Donald Trump.
[03:12.410 --> 03:16.250]  The nominee says he won't accept Ted Cruz's endorsement.
[03:17.090 --> 03:20.770]  You know, it's all about what's going on in the Republican Party.
[03:21.030 --> 03:21.910]  Nazi salute.
[03:22.650 --> 03:30.710]  Some of these don't seem like the things that are most popular, but perhaps the things that are most interesting to the team of curators.
[03:30.710 --> 03:34.150]  So how do they come up with this list of what's trending?
[03:34.870 --> 03:38.110]  Well, Facebook has an answer for this in their handy help section.
[03:39.030 --> 03:42.070]  Trending shows you topics that have recently become popular.
[03:42.590 --> 03:49.970]  The topics you've seen are based on factors including engagement, timeliness, pages you've liked, and your location.
[03:50.410 --> 03:51.970]  If only that were true.
[03:52.210 --> 03:57.910]  If that were true, then perhaps we really would have a neutral-ish Facebook.
[03:57.910 --> 04:04.770]  Instead, what we have are manually curated trends.
[04:05.430 --> 04:15.390]  Can I get a show of hands of folks in this room who's kind of familiar with the story of how Facebook was suppressing conservative news or allegedly suppressing conservative news?
[04:15.750 --> 04:17.690]  Okay, so that's some of you, but not everyone.
[04:17.690 --> 04:27.510]  So to bring you all up to speed, this trending section that you see here, that's actually not algorithmic.
[04:27.750 --> 04:30.410]  That is curated by people.
[04:30.550 --> 04:34.550]  People like you and me who are not employees but contractors of Facebook.
[04:35.490 --> 04:40.590]  One or more of them complained to a news organization about their work conditions.
[04:40.590 --> 04:45.270]  And a subset of that complaint was that conservative news was being suppressed.
[04:45.930 --> 04:48.210]  That seems possible.
[04:48.610 --> 04:56.090]  Whether or not it's actually true, the fact does remain that this is human curated.
[04:56.450 --> 04:58.170]  There is inherent bias.
[04:58.410 --> 05:00.250]  It is not what's most popular.
[05:00.250 --> 05:05.250]  It's what the curation team deems to be worth viewing, which of course inherently is biased.
[05:05.250 --> 05:11.030]  This is what their tool looks like, their review tool for deciding what goes up or down.
[05:11.250 --> 05:13.990]  And there's a 28-page document that's been released.
[05:14.010 --> 05:21.910]  It's an internal Facebook document, and it talks about all of their policies and procedures and blacklisting of this topic or that topic.
[05:23.050 --> 05:26.670]  Gizmodo headline, we routinely suppress conservative news.
[05:27.710 --> 05:33.510]  Depending on who was on shift, things would be blacklisted or trending, says a former curator.
[05:33.510 --> 05:35.890]  That's the danger with curation, right?
[05:36.050 --> 05:37.230]  Like, there are humans doing it.
[05:37.310 --> 05:39.030]  How can it possibly be free of bias?
[05:40.710 --> 05:46.350]  The company does state that it's neutral, but the leadership isn't neutral.
[05:47.750 --> 05:52.430]  Here's one headline, Mark Zuckerberg throws shade at Donald Trump during his speech.
[05:53.030 --> 05:58.130]  He did incite Trump by name, but it's clear his affiliations lie.
[05:59.330 --> 06:02.970]  And it's clear where the affiliations of the employees lie.
[06:04.030 --> 06:10.010]  They went so far as to say in an internal meeting, can we prevent President Trump?
[06:10.210 --> 06:12.790]  And they also asked, are we obligated to?
[06:13.310 --> 06:14.530]  Which is interesting.
[06:15.830 --> 06:18.330]  If you could make sure all phones are off, please.
[06:23.700 --> 06:25.660]  It's dark, we could have a dance party.
[06:25.840 --> 06:27.300]  We can do whatever we want.
[06:27.580 --> 06:28.600]  I have the stage.
[06:29.740 --> 06:30.540]  I'm biased.
[06:30.840 --> 06:31.600]  I'll do what I want.
[06:32.800 --> 06:34.160]  With your consent though, right?
[06:34.300 --> 06:34.820]  This is hope.
[06:35.000 --> 06:36.800]  It should be all of us together.
[06:37.280 --> 06:38.640]  One, ideally.
[06:38.640 --> 06:58.300]  So, this was very interesting to me that the Facebook employees asked not what their ethical responsibilities were as a company or as a supposedly neutral company, but they asked if they were obligated to prevent one candidate from becoming successful.
[06:59.500 --> 07:01.480]  And (((Mark Zuckerberg))) said, no.
[07:01.660 --> 07:05.300]  No, we are not going to officially prevent a President Trump.
[07:05.600 --> 07:16.480]  But still, how much is the company controlling the employees and how much are they able to influence in the ways that they feel are ethical?
[07:17.040 --> 07:24.920]  One thing we can say is that the company is officially neutral and the people inside of it are all over the place.
[07:25.960 --> 07:33.860]  Board member (((Peter Thiel))) is now famously a delegate for Trump and spoke at the RNC.
[07:34.400 --> 07:44.040]  So, one thing we can rely on within many of these companies is even the ones who are leaning heavily in one direction or another, it's still a mixed bag of people.
[07:46.560 --> 07:50.080]  It would be legal if Facebook wanted to impact the elections.
[07:50.080 --> 07:53.780]  The same way it's legal for the New York Times or the Washington Post.
[07:54.880 --> 07:57.860]  Facebook does not have an obligation to neutrality.
[07:58.000 --> 08:01.760]  It's a business decision that they are making that they want to do things in this way.
[08:02.040 --> 08:04.400]  It is arguably part of free speech.
[08:04.600 --> 08:08.940]  They do need to stay within campaign finance regulations, but that is not so difficult to do.
[08:09.940 --> 08:18.120]  According to UCLA law professor Eugene Volokh, who told Gizmodo, Facebook has the same First Amendment rights as the New York Times.
[08:19.000 --> 08:21.960]  You could argue that Google does not have these rights.
[08:22.140 --> 08:26.820]  That Google has an obligation to provide the most accurate or highest quality search results.
[08:27.280 --> 08:31.240]  This law professor argues that Google does not have that obligation.
[08:31.900 --> 08:33.360]  You could argue either way.
[08:33.540 --> 08:35.760]  I think either perspective is valid.
[08:36.120 --> 08:37.600]  I'm not a law expert though.
[08:38.660 --> 08:40.940]  As for Twitter, I don't know.
[08:41.080 --> 08:41.980]  Is it business?
[08:42.000 --> 08:42.780]  Is it bias?
[08:42.960 --> 08:44.500]  Or is it just a shit show?
[08:45.160 --> 08:46.120]  I don't know.
[08:46.440 --> 08:47.800]  It's a good question.
[08:48.840 --> 08:50.760]  I think about this a lot, I wonder.
[08:52.560 --> 08:54.080]  Let's look at their curation.
[08:54.880 --> 09:01.740]  So this is a sample of the moments that were on Twitter today, earlier today.
[09:02.260 --> 09:06.080]  It includes funniest GIFs and memes to come from the RNC.
[09:08.060 --> 09:11.380]  Donald Trump is now the official nominee.
[09:12.320 --> 09:15.120]  Gayest RNC, but what does that mean?
[09:16.000 --> 09:37.540]  It seems like whoever curated this was looking for the most entertaining, not necessarily the most newsworthy or the most tweeted, compared to this set of moments, which is fairly dry, I think, but is supposed to be what was the top three tweeted moments.
[09:37.800 --> 09:43.560]  This came from the Twitter government account as opposed to the moments account.
[09:43.820 --> 09:46.620]  They're both Twitter curated accounts.
[09:46.620 --> 09:58.600]  And I think they're interesting as examples of curation, which just, much like Facebook, Twitter moments, is curated by people, people who have biases.
[09:59.620 --> 10:05.440]  And the lack of transparency around how all of this is done leads to questions and skepticism.
[10:08.700 --> 10:14.660]  Why I'm not voting for Hillary was the number one trending hashtag, and then it wasn't.
[10:14.780 --> 10:18.500]  It was replaced by another hashtag, Hillary 2016.
[10:19.420 --> 10:30.880]  Now, it's possible that one hashtag just overtook the other, but Twitter's lack of transparency leads to skepticism and questions and just a general lack of trust.
[10:31.900 --> 10:34.580]  And I wonder whether bans will impact the election.
[10:34.960 --> 10:40.340]  Twitter is banning conservatives who, in some cases, may very rightfully deserve it.
[10:41.860 --> 10:49.080]  But nonetheless, it could potentially impact the election if those folks are no longer tweeting.
[10:49.840 --> 11:00.480]  You could say that's good or you could say that's bad, but it does seem clear that Twitter's bans are more heavily conservative than in other directions.
[11:01.140 --> 11:04.340]  I'm not saying whether that's good or bad, but I do think it would impact the election.
[11:05.400 --> 11:07.700]  So let's shift gears a bit and talk about Google.
[11:08.840 --> 11:14.820]  Apparently, it wasn't a glitch when a State Department video was cut from YouTube.
[11:15.520 --> 11:17.100]  This is very interesting.
[11:18.640 --> 11:19.880]  It's very interesting.
[11:20.740 --> 11:25.500]  Classified emails suggest that Hillary Clinton worked with Google and YouTube to block those videos.
[11:25.800 --> 11:31.100]  And we do know that Google and the White House are very, very friendly.
[11:32.480 --> 11:33.980]  Orr Books is here.
[11:34.200 --> 11:47.320]  And they're vending this book, which is an interesting read on the topic, from Julian Assange, which documents his conversations with Eric Schmidt, which documents Eric Schmidt's closeness with the White House.
[11:50.830 --> 12:03.730]  Eric Schmidt started a startup which has the aim of helping the Clinton campaign, which actually, you know, it's like pretty cool the way that campaigns are becoming much more technical.
[12:05.170 --> 12:08.110]  But not a lot of transparency there either.
[12:08.430 --> 12:15.810]  This is the webpage for the groundwork, the startup started by Eric Schmidt, that is working closely with the Clinton campaign.
[12:16.090 --> 12:17.070]  It's so stealthy.
[12:17.310 --> 12:18.390]  That's all that's there.
[12:18.970 --> 12:19.750]  That's it.
[12:19.930 --> 12:20.650]  That's everything.
[12:20.650 --> 12:30.210]  The only reason why I have any background information on the groundwork is because it's been documented elsewhere by reporters and researchers.
[12:30.910 --> 12:36.050]  And I'm actually quite fond of how the Clinton campaign has handled many things.
[12:36.190 --> 12:40.310]  I think that they seem to be a very intelligent and well-run campaign.
[12:40.670 --> 12:46.290]  Nonetheless, it is noteworthy that many of their top people do hail from Google.
[12:46.290 --> 12:49.390]  There's some kind of a close relationship there.
[12:50.790 --> 12:52.690]  That said, search does seem legit.
[12:53.170 --> 13:00.310]  Questions were raised about whether Google's search results were sanitized on behalf of Hillary Clinton.
[13:00.730 --> 13:03.670]  And this looked as if it was some kind of a scandal.
[13:04.510 --> 13:14.430]  I mean, if you look at the results, they're not as interesting as the results you get with the same query on Bing or on other search engines.
[13:14.430 --> 13:16.670]  But the same is true for Trump.
[13:17.410 --> 13:22.330]  And in fact, Google autocomplete is dull like for everyone.
[13:23.750 --> 13:26.610]  You know, name a criminal, name your favorite criminal.
[13:26.890 --> 13:29.910]  And type your favorite criminal's name into Google.
[13:30.090 --> 13:32.350]  And they just won't autocomplete crime or criminal.
[13:33.430 --> 13:35.330]  These algorithms are strange.
[13:35.950 --> 13:42.290]  Google doesn't want you searching for a person's name and then the word criminal for whatever reason.
[13:42.290 --> 13:47.530]  Maybe someone who worked on the algorithm decided that, like, those just weren't great search results.
[13:48.850 --> 14:02.130]  I think if there's any one theme that comes out from this talk by me today, it's that this lack of transparency raises questions and is frustrating and confusing.
[14:02.130 --> 14:05.730]  I mean, we really don't know the reason why Google works this way.
[14:06.430 --> 14:13.510]  And because we don't know why, sometimes we imagine or create theories about why that aren't very favorable.
[14:14.170 --> 14:20.150]  And the truth is we actually don't know whether this is secretly to help one group or another group.
[14:20.150 --> 14:21.190]  We just don't know the reasons.
[14:21.390 --> 14:22.110]  Help the criminals.
[14:22.970 --> 14:23.930]  What did you say?
[14:24.170 --> 14:25.070]  Help the criminals.
[14:25.510 --> 14:26.030]  Maybe.
[14:26.410 --> 14:27.130]  To conceal them.
[14:27.230 --> 14:29.130]  Maybe it's to help conceal criminals.
[14:29.230 --> 14:29.690]  Who knows?
[14:29.970 --> 14:32.490]  The lack of transparency for me is extremely frustrating.
[14:32.750 --> 14:39.210]  But it does seem clear that it's not specifically to help, say, Clinton or Trump.
[14:40.350 --> 14:42.490]  Google autocomplete, dull for everyone.
[14:44.030 --> 14:45.710]  Let's go back to Facebook.
[14:47.390 --> 14:51.410]  This actually makes me a little worried and nervous and sad.
[14:52.650 --> 14:55.150]  Facebook runs experiments all the time.
[14:55.710 --> 14:57.810]  They don't always tell us.
[14:58.310 --> 14:59.750]  They don't have to.
[15:02.010 --> 15:04.870]  Google messed with your feed in 2012.
[15:05.910 --> 15:07.590]  That's not the only time.
[15:08.090 --> 15:09.930]  They did that to promote voting.
[15:10.670 --> 15:13.510]  And they also messed with our feed in 2010.
[15:13.910 --> 15:15.230]  Which is their right, right?
[15:15.370 --> 15:17.010]  Like, they get to do whatever they want.
[15:17.410 --> 15:19.270]  This is their stage.
[15:19.850 --> 15:21.430]  They are in charge.
[15:21.750 --> 15:24.910]  I think the issue with Facebook is it feels like it's my page.
[15:25.630 --> 15:29.730]  But it's very heavily curated what they're going to show for my friends.
[15:30.110 --> 15:35.390]  And then they've started to really do those messages at the top of the page.
[15:35.390 --> 15:38.650]  Like, suggesting that someone goes out and votes.
[15:40.430 --> 15:42.630]  They did this to 60 million users.
[15:42.990 --> 15:47.230]  That's a lot of users that actually really will impact an election.
[15:48.510 --> 15:51.630]  This was the paper that they published about it.
[15:51.790 --> 15:53.650]  So at least they came out and they wrote about it.
[15:53.770 --> 15:54.630]  But it was after the fact.
[15:56.010 --> 16:05.570]  Here we report results from a randomized controlled trial of political mobilization messages delivered to 61 million Facebook users during the congressional elections.
[16:09.080 --> 16:14.540]  So it's a kind of digital gerrymandering if they were to show this to some people and not to others.
[16:14.740 --> 16:16.200]  And that actually was the experiment.
[16:16.200 --> 16:22.140]  Some people got shown buttons that were more encouraging to vote than other people.
[16:23.020 --> 16:26.900]  Now, we could just say civic engagement is good and that it's neutral.
[16:27.560 --> 16:30.780]  And for the most part, it's hard to argue against that.
[16:32.740 --> 16:45.860]  But when I spoke to someone who works in Democratic campaigns and I said, I would like to be helpful this election cycle to the Democrats, you know, how can I, what can I do?
[16:45.860 --> 16:46.460]  I'm here.
[16:46.660 --> 16:47.360]  I'm here for you.
[16:47.940 --> 16:53.160]  And they said that they didn't want me to promote their specific candidate.
[16:53.920 --> 17:01.020]  That the most helpful thing I could do for the Democratic Party would be to just register more voters.
[17:01.820 --> 17:03.980]  Which was strange to me.
[17:04.540 --> 17:09.760]  I thought that the way to help a candidate win was to help the candidate win.
[17:11.000 --> 17:15.680]  And they said, oh, Alyssa, you know, you've worked on Republican races.
[17:15.940 --> 17:17.060]  Because I'd helped Ron Paul.
[17:17.720 --> 17:21.220]  But Democrats, Democrats, we run races by registering voters.
[17:21.400 --> 17:22.080]  That's how we win.
[17:22.400 --> 17:29.920]  So, you know, when I see companies working to get out the vote, I'm a little bit skeptical.
[17:30.520 --> 17:33.740]  Because Democrats win elections by increasing voter turnout.
[17:33.740 --> 17:39.920]  So, if a platform like Facebook is working to increase voter turnout, like on the one hand, of course, that's great.
[17:40.040 --> 17:42.400]  On the other hand, is it neutral?
[17:42.520 --> 17:43.080]  Maybe not.
[17:44.860 --> 17:47.140]  And is neutrality an impossible goal?
[17:47.540 --> 17:49.120]  I don't, I don't know.
[17:49.400 --> 17:51.180]  Here's another curation.
[17:52.660 --> 17:54.100]  Escape his caricature.
[17:54.560 --> 17:56.280]  Donald Trump passes up the chance.
[17:57.480 --> 17:59.040]  It's, it's just not neutral.
[18:01.580 --> 18:04.160]  And something much more favorable to Clinton today.
[18:09.440 --> 18:13.160]  So, the next place I take that is this very hope-like critique.
[18:13.440 --> 18:28.180]  Where I think the chance that we have for having actual neutrality on our platforms is if we remove these centralized curators and can actually just exchange one-on-one.
[18:28.180 --> 18:32.880]  The censorship is problematic because there's just no transparency.
[18:33.280 --> 18:35.160]  We just really don't know what's going on.
[18:37.020 --> 18:39.740]  Can I start opening it up for questions?
[18:40.040 --> 18:42.500]  Although, it's very dark, so I can't see anyone.
[18:42.680 --> 18:43.520]  I might need some help.
[18:45.600 --> 18:48.760]  Could I get some help from one of the conference organizers?
[18:49.380 --> 18:49.860]  Thanks.
[18:53.870 --> 18:56.190]  Alright, let's open it up for questions.
[19:07.770 --> 19:09.690]  If you have a question, the mic.
[19:10.370 --> 19:10.950]  Oh, okay.
[19:11.110 --> 19:11.210]  Thank you.
[19:11.230 --> 19:12.750]  The mic is right down here.
[19:13.030 --> 19:13.870]  I can see it now.
[19:14.050 --> 19:15.210]  I'm adjusting to the lights.
[19:15.590 --> 19:16.150]  Thank you.
[19:16.250 --> 19:16.490]  Awesome.
[19:17.130 --> 19:17.770]  Yes, please.
[19:18.670 --> 19:25.230]  So, I have worked for Google and for Twitter, but these are in no way representing views of either Google or Twitter.
[19:25.550 --> 19:25.950]  Okay.
[19:26.150 --> 19:26.850]  Positive negatively.
[19:26.850 --> 19:34.450]  So, I think you're making a couple of logical mistakes here in your analysis.
[19:34.990 --> 19:39.370]  One is, you're ignoring the echo chamber effect.
[19:39.850 --> 19:49.650]  When I'm looking at Facebook pages, and what Facebook would show me, and is interesting for me, is what I have said and what my friends have said.
[19:49.650 --> 19:56.510]  Now, most of my friends are not hardcore Republicans who would vote for Trump.
[19:57.230 --> 20:00.750]  So, it is natural, except this guy here.
[20:01.610 --> 20:05.850]  So, it's natural that Facebook won't show me stuff about Trump.
[20:05.850 --> 20:22.290]  Now, if I were to just create a sock puppet that would listen to the Westboro Baptist Church, and to Rudy Giuliani, and Trump, and other such things, maybe I would get a different set of results.
[20:22.290 --> 20:35.690]  Have you tried that experiment of creating a persona that does not fit the model you're trying to accuse, or the model that you're suggesting that Facebook is promoting, and seeing what the results you get are?
[20:37.370 --> 20:38.510]  I think that's right.
[20:38.670 --> 20:44.150]  I think that it would be helpful to see it from different vantage points.
[20:44.230 --> 20:44.570]  Right, but before...
[20:44.570 --> 20:46.350]  I've solicited some of that.
[20:46.930 --> 20:51.590]  It doesn't change that the moments, for the most part, are universal.
[20:52.670 --> 21:02.930]  I don't... that's true for Twitter, and I believe it's true for Facebook as well, that these aren't as highly customized as we would think.
[21:04.190 --> 21:09.570]  Facebook doesn't show you... Facebook doesn't show everyone the same thing.
[21:11.390 --> 21:17.130]  It just never... you know, there's many things that people just don't see.
[21:18.670 --> 21:20.490]  That's true for your news feed.
[21:21.330 --> 21:25.790]  I'm curious about how true it is for what's trending at large.
[21:26.690 --> 21:28.270]  Can you come to the mic, please?
[21:29.050 --> 21:30.710]  I'm not done, but go ahead.
[21:31.430 --> 21:41.270]  That doesn't... that doesn't apply to... what it applies to is each person's news feed.
[21:41.510 --> 21:41.990]  Right.
[21:42.410 --> 21:44.630]  That is different for every person.
[21:45.050 --> 21:45.570]  That's right.
[21:46.690 --> 21:56.630]  What I'm referring to here is... the trending section... here.
[21:59.310 --> 22:01.170]  Yes, the news feed is very customized.
[22:04.500 --> 22:15.600]  Well, the news feed will be customized for the sort of articles I would be likely to read, so I would click on them, look at the ads, and then make Facebook or Google or Twitter money.
[22:15.600 --> 22:25.160]  They wouldn't show... if they showed me an ad about, oh, I don't know, the starving children of Africa, I would probably not click on it and not read the article.
[22:25.520 --> 22:27.240]  And, you know, not get them money.
[22:27.360 --> 22:28.360]  So, of course, they won't do that.
[22:28.520 --> 22:30.620]  They do a very good job of customizing those.
[22:30.620 --> 22:31.140]  Yeah.
[22:31.760 --> 22:53.080]  So, how can you differentiate... what is the... your algorithm for differentiating between a feed that is evilly curated to only show those evil people that you don't want... that you're accusing of... sorry, let me tone it down.
[22:53.080 --> 23:09.600]  You know, how can you differentiate between a feed that is customized according to some guidelines from above versus a feed that happens to appear that way because of all your prior posting history?
[23:09.600 --> 23:24.460]  Well, we need to separate news feed, which is highly personalized and highly algorithmic versus what you can see here, which is like national trending topics, which is curated by humans.
[23:25.580 --> 23:37.100]  This curation that you're talking about, the people who have... you mentioned that, oh, ex-curator said that, and the rumor said that.
[23:37.320 --> 23:43.000]  How have you actually verified that this is indeed the case?
[23:43.180 --> 23:50.840]  You know, I heard from my dog yesterday that, you know, Mark Zuckerberg did something.
[23:50.840 --> 24:02.940]  Like, just because you said, an ex-curator said that, did you, you know, whoever published that story, did they go out and actually see that this was actually an ex-curator?
[24:03.280 --> 24:15.420]  You know, the journalistic standards that we used to hold up, the Washington Post and the New York Times 100 years ago or 50 years ago, are a lot different than...
[24:15.420 --> 24:23.740]  And we expect when the New York Times or the Washington Post says, you know, something happened, that the journalists have actually fact-checked it.
[24:23.740 --> 24:31.840]  I think that that's fair, that we just don't know what's going on with the curation and the trends.
[24:32.120 --> 24:38.940]  And I certainly don't mean to say that we know for sure that there's this slant or that slant.
[24:38.940 --> 25:01.500]  I think that it's compelling, the evidence that there was...the evidence that what was going on in this team that was a contract team that wasn't being given a lot of attention, that, like, there may have been some things happening there that weren't up to Facebook standards or up to the standards that we would want.
[25:01.500 --> 25:03.580]  But that's a lot of the point.
[25:03.600 --> 25:08.740]  That's a lot of the point of this talk is that so much of this is a black box and we just don't know.
[25:09.020 --> 25:15.240]  And that it would be possible for social media companies to have an outsized impact on the elections.
[25:15.400 --> 25:21.700]  And in the absence of transparency about their methods, we're left not knowing.
[25:23.680 --> 25:24.540]  Thank you.
[25:25.380 --> 25:26.460]  Next speaker, please.
[25:26.580 --> 25:28.040]  We'll give up the mic and come back.
[25:28.720 --> 25:29.980]  Thank you for your question.
[25:30.360 --> 25:31.180]  And thank you for the cheering.
[25:32.900 --> 25:33.300]  Hi.
[25:33.440 --> 25:37.440]  I was particularly interested in your last slide where you had a P2P model about kind of...
[25:37.440 --> 25:37.680]  Yes.
[25:37.920 --> 25:44.340]  ...propogating what is most interesting to, you know, like, kind of figuring out in a decentralized way what is most interesting kind of to show people.
[25:44.880 --> 25:51.620]  I do kind of wonder how that...how you would defend against creating echo chambers or just little pods of people just share the same news amongst each other.
[25:52.300 --> 26:02.200]  And kind of all this thing of that, I was thinking, you know, there's a lot of talk about how everyone getting news from a different source kind of, like, helps to just make just many sides that just don't even want to talk to each other anymore.
[26:02.200 --> 26:13.800]  And how, like, having fewer news sources, like, everyone would watch one or two of them and there'd be, like, more common watching and so you would have less extreme, I guess, divisiveness in kind of the whole political process.
[26:14.620 --> 26:25.660]  So, I don't know, do you think there's a kind of a part to play for these outlets to maybe show multiple, like, a diversity of things that may be interesting so they aren't just making echo chambers?
[26:26.420 --> 26:27.720]  I guess it's kind of two questions.
[26:27.780 --> 26:36.120]  So, your question, if I understand it, is how would we do a peer-to-peer model that would bypass the echo chamber problem?
[26:36.220 --> 26:36.660]  Yeah, sure.
[26:36.680 --> 26:37.220]  Is that quite right?
[26:37.480 --> 26:37.600]  Yeah.
[26:38.600 --> 26:40.080]  That's a really hard problem.
[26:40.340 --> 26:41.680]  That's like a human problem, right?
[26:41.680 --> 26:44.160]  I think you solved that in a human way.
[26:45.180 --> 26:58.260]  I have a lot of friends who have different points of view and I engage with them and I've learned to stop doing that as much because I've gotten negative feedback.
[26:58.920 --> 27:05.880]  Because when I engage with people at very different points of view, sometimes they will troll me or sometimes they'll be very mean or they'll troll my friends.
[27:05.880 --> 27:13.120]  Or people who have a similar point of view to me will get mad at me for interacting with people who they think are wrong.
[27:13.640 --> 27:22.320]  So, I actually started out as someone who was mentored by Ethan Zuckerman, who wrote Rewire.
[27:22.320 --> 27:26.820]  Like his old thesis was that the Internet is an echo chamber and we need to break out of that.
[27:27.400 --> 27:32.620]  So, like I came into the Internet with that mentality, being mentored by him.
[27:32.780 --> 27:41.880]  And now I'm becoming smaller because of all the bad feedback that I'm getting as I actually engage with diverse views.
[27:41.880 --> 27:43.740]  And I say that in a sheepish way.
[27:43.840 --> 27:46.940]  I don't want to... I don't want my worldview to become smaller.
[27:46.940 --> 27:48.720]  I don't want to become smaller as a person.
[27:49.300 --> 27:50.520]  But it's what's happening.
[27:50.820 --> 27:51.060]  Sure.
[27:51.240 --> 27:56.060]  I started deleting posts on my Facebook last week for the first time.
[27:56.280 --> 27:58.500]  Kind of apologizing to people and explaining why.
[27:59.660 --> 28:03.480]  So, I think one first step is for us to be nicer to each other.
[28:03.600 --> 28:03.920]  Sure.
[28:04.100 --> 28:05.220]  To agree to disagree.
[28:06.300 --> 28:06.740]  Yeah.
[28:06.740 --> 28:12.860]  Like, if we could agree to disagree, then it would be a lot easier to have free exchange.
[28:14.180 --> 28:14.580]  Yeah.
[28:14.760 --> 28:18.900]  I think the bottom line for me is that's like a human problem and a kindness problem.
[28:19.600 --> 28:26.080]  And like, can we be interested in views beyond our own with the humbleness that there's something to learn problem?
[28:26.440 --> 28:26.840]  Sure.
[28:26.980 --> 28:28.620]  So, that's a really big ask, which is awesome.
[28:28.740 --> 28:29.060]  And I love it.
[28:29.160 --> 28:29.740]  I think it should really happen.
[28:29.740 --> 28:38.460]  In the meantime, maybe, do you think it's possible for, like, Twitter and Facebook to use their kind of power that they have just by doing these feeds?
[28:38.760 --> 28:42.360]  Not feeds, but like the trending topics, to try and force that hand a little bit?
[28:42.480 --> 28:45.400]  Like, do you think they have the capacity to be more even-handed, I guess?
[28:45.520 --> 28:50.140]  Can they all just be like the BBC or something like that where it's dry but it tends to be a little more neutral?
[28:50.340 --> 28:54.380]  I think that Twitter and Facebook just have tremendous power right now.
[28:54.380 --> 28:54.580]  Sure.
[28:54.780 --> 28:54.940]  Oh, yeah.
[28:55.100 --> 28:55.320]  Absolutely.
[28:55.880 --> 29:14.120]  And part of what I came away with as I did the research for the talk was just the respect for how much power Twitter and Facebook and, you know, Instagram and Snapchat have for getting us to take actions and for changing our lens on world events.
[29:14.340 --> 29:14.460]  Right.
[29:14.880 --> 29:17.900]  And that's what's frustrating to me about the notion that they're neutral.
[29:17.900 --> 29:19.160]  They're just so powerful.
[29:19.440 --> 29:22.440]  And I want to see more acknowledgement of that.
[29:23.000 --> 29:33.240]  And as for, like, how they could use that for good, I think the first step might be for them to kind of acknowledge that there are entities that are trying to do things in the world.
[29:33.240 --> 29:40.060]  As long as Facebook and Twitter take a stand and say we're neutral and we're not really doing much.
[29:40.060 --> 29:46.060]  But by the way, we're going to run this experiment and some of you will vote and some of you won't based on what we show you.
[29:46.780 --> 29:56.080]  As long as there's this pretense that they're like bystanders, I think it would be hard for them to become best actors.
[29:56.360 --> 29:56.580]  Right.
[29:56.780 --> 29:56.880]  Sure.
[29:57.060 --> 29:57.320]  Thank you.
[29:57.860 --> 29:58.500]  Thank you.
[30:00.880 --> 30:01.260]  Hi.
[30:01.400 --> 30:01.780]  Thank you.
[30:03.000 --> 30:11.660]  I probably should confess before I start is that I'm consciously not on Facebook because I am dismayed about the amount of power they have.
[30:12.360 --> 30:14.260]  But what I'm saying is...
[30:14.260 --> 30:15.420]  You're in good company here.
[30:15.840 --> 30:21.280]  What I'd like to point out is that no private company has transparency unless they want to.
[30:21.420 --> 30:23.980]  I mean, just think about the credit reporting agencies.
[30:23.980 --> 30:32.480]  And the problem that I have with this Facebook story is they have no obligation to be neutral.
[30:32.740 --> 30:43.560]  And the truth is if you're getting your news and you're allowing companies like Facebook or Twitter or Snapchat to mold your political views, there's something wrong right there.
[30:44.060 --> 30:46.320]  You need to take it back from these companies.
[30:46.440 --> 30:48.460]  They have only as much power as we give to them.
[30:48.460 --> 30:55.040]  And why you're giving these companies the power to mold your political views is be on me.
[30:55.640 --> 30:56.420]  I think...yeah.
[30:56.820 --> 30:57.380]  Yeah.
[30:58.500 --> 31:00.240]  I think that's a good lesson.
[31:00.620 --> 31:11.140]  I think that was my takeaway from Snowden's revelations from so many other moments where we've seen these companies aggregating data and becoming very powerful.
[31:11.340 --> 31:13.340]  That's power that they only have that we give them.
[31:13.500 --> 31:16.340]  So that's a very important point.
[31:16.340 --> 31:21.920]  I would argue that it's my hope that companies will be true to their word.
[31:22.620 --> 31:28.180]  And it's my hope that companies will be true to the power and responsibility they have.
[31:28.460 --> 31:30.080]  That's a very optimistic view.
[31:30.200 --> 31:31.660]  They don't need to do that.
[31:32.400 --> 31:38.560]  But we can call on them and, you know, come together and say, we think it matters.
[31:39.540 --> 31:42.240]  Or we can, you know, I admire your stance.
[31:46.560 --> 31:47.000]  Hi.
[31:49.940 --> 31:50.380]  Okay.
[31:51.140 --> 31:51.420]  Thanks.
[31:51.680 --> 31:52.620]  Short people unite.
[31:52.960 --> 31:53.400]  Anyway.
[31:54.420 --> 31:56.780]  I'm sitting because I'm like five feet tall.
[31:58.160 --> 32:02.040]  Actually, the previous speaker actually had one of the things I wanted to say.
[32:02.200 --> 32:10.560]  But the other thing I want to say is, I mean, I know a lot of people are suspicious of government because, you know, you know, there are all these, well, I don't have to go into why.
[32:10.560 --> 32:19.800]  But I'm just thinking that I think we should also be suspicious of private business because, you know, that's what they are and that people are just way too trusting.
[32:19.800 --> 32:28.340]  And I don't, I really don't think it's reasonable to expect that these companies are going to, you know, be what we want them to be.
[32:28.500 --> 32:37.560]  I mean, we have to decide, you know, like she said, you know, decide to engage or not, you know, but with our eyes open, they're in it to make money, they can do what they want.
[32:37.560 --> 32:41.700]  And just, you know, that's basically where it's at.
[32:41.820 --> 32:43.100]  They're in it to make money.
[32:43.220 --> 32:57.380]  And in some cases, I believe that technologists at these companies come in either with like a genuinely ethical point of view from their point of view, you know, like, is it our obligation to stop Trump?
[32:57.540 --> 32:58.980]  That's an employee question.
[32:58.980 --> 33:00.020]  That's not about money.
[33:00.180 --> 33:02.540]  Well, that's kind of about civilization, I guess.
[33:02.540 --> 33:07.140]  That's probably one of those core value things that people just do because of who they are, you know?
[33:07.380 --> 33:12.280]  Well, and when you look at people running experiments, that's coming out of curiosity.
[33:12.980 --> 33:14.360]  But with a certain disregard...
[33:14.360 --> 33:19.980]  Well, I think experiment, but the thing is that when it's a private company doing experiments, it's a kind of a different...
[33:19.980 --> 33:20.940]  But I guess that's the thing.
[33:21.500 --> 33:22.680]  People know about it now.
[33:23.020 --> 33:31.900]  I guess the only other thing I would mention is back in the day when we used to have like comment sections that were not moderated, not curated, any of that stuff.
[33:31.900 --> 33:44.600]  One of the things that happens a lot is that you get the trollish types, you get wall-to-wall shit, and then everybody leaves, everybody leaves, and you got nothing.
[33:45.200 --> 33:47.620]  Well, maybe we do smaller groups.
[33:47.800 --> 33:54.540]  It's an interesting question and one that I hope people will continue of what else we could be doing.
[33:54.840 --> 33:55.520]  Anyway, okay.
[34:01.130 --> 34:01.610]  Hi.
[34:01.870 --> 34:02.210]  Hi.
[34:02.490 --> 34:05.450]  This is not like really more of a question, it's more of a comment.
[34:05.450 --> 34:05.790]  Please.
[34:05.790 --> 34:08.570]  And I want to build off of what the first man was speaking about.
[34:08.810 --> 34:11.310]  But I want to give more of a deferring outlook.
[34:11.710 --> 34:17.930]  He was saying that maybe if you do follow conservative pages, things like that, maybe you do get different search results.
[34:18.190 --> 34:20.710]  And I myself, I support Donald Trump.
[34:20.710 --> 34:26.850]  I follow the GOP on Twitter, Trump, and the other candidates when they were still competing in the primaries.
[34:27.570 --> 34:33.970]  And my hashtag, the trends, everything is still filled with Trump hate, things like that.
[34:34.510 --> 34:37.730]  And so I really, and I get Hillary ads all the time.
[34:37.970 --> 34:44.510]  I really don't think it has anything to do with who you follow or what you say, because I'm always tweeting pro-conservative things out on Twitter.
[34:46.110 --> 34:53.650]  And I think that even if it was based off of a majority of what people are tweeting about, I still think I disagree.
[34:54.150 --> 35:00.530]  Because I don't think, if Trump would have won the primaries by such a landslide, he obviously has such a large following.
[35:00.530 --> 35:05.410]  And I don't think that the majority of people tweeting would always be tweeting such awful things about him.
[35:06.230 --> 35:11.610]  Like, I barely ever see if I've ever seen a positive hashtag about him in the trends.
[35:11.850 --> 35:19.190]  So I really do think that the large social medias are trying to turn people left leaning.
[35:19.390 --> 35:20.010]  That's my opinion.
[35:20.750 --> 35:27.410]  Twitter in particular, there's more evidence that there is an anti-conservative bias.
[35:27.410 --> 35:30.990]  And that shows up primarily in who they choose to ban.
[35:31.190 --> 35:36.350]  And I'm certainly not defending that because they ban people who are trolling people.
[35:37.630 --> 35:42.830]  But I would argue that I haven't seen the same number of bans on the other side.
[35:44.570 --> 35:49.530]  In other social media organizations, I believe personally that that's less clear.
[35:52.610 --> 36:00.290]  But I do believe that there are legitimate concerns about bias at Twitter.
[36:00.490 --> 36:03.550]  Twitter has made less statements than Facebook about being neutral.
[36:03.550 --> 36:08.050]  And if I could just, like, the Trump Yourself campaign that Hillary Clinton started.
[36:09.130 --> 36:14.170]  If you actually go to her website and look in that thing that she just started.
[36:14.470 --> 36:20.130]  It says how he doesn't want to support lesbian and gay marriage.
[36:20.410 --> 36:23.890]  But, like, I've seen him speak live in speeches before.
[36:24.130 --> 36:27.370]  He said that he wants to protect people like that from hateful ideologies.
[36:27.750 --> 36:29.630]  Foreign ideologies, not domestic.
[36:29.630 --> 36:30.810]  Even foreign.
[36:30.990 --> 36:37.190]  There are such a large number of people outside the United States that still have such barbaric points of views.
[36:37.210 --> 36:44.670]  Perhaps he's calling those ideals, you know, those ideas of the domestic as foreign themselves.
[36:45.090 --> 36:46.490]  I don't think Trump actually...
[36:46.490 --> 36:51.110]  Well, we should probably move to next question lest we get off track.
[36:51.170 --> 36:56.970]  But it's helpful to hear that your trends don't seem to reflect your ideology.
[36:57.190 --> 36:57.650]  That's helpful.
[37:00.490 --> 37:12.250]  I don't know that until this current kerfuffle with Facebook's trending news stories that either Facebook or Twitter or any of these companies professed that they're...
[37:12.250 --> 37:19.610]  Or, like, really made a big deal out of, like, the neutrality of the topics that they're showing in their trending sections, right?
[37:20.170 --> 37:22.650]  Neutrality wasn't as important when there was less curation.
[37:22.650 --> 37:23.790]  Well, but I think...
[37:24.850 --> 37:25.330]  I'm...
[37:25.990 --> 37:29.950]  I think that what people often mistook...
[37:29.950 --> 37:34.870]  Or what people are mistaking for neutrality is, like, oh, that's 100% algorithm powered.
[37:35.110 --> 37:36.890]  No human ever touches it.
[37:37.810 --> 37:39.650]  And I think that you would be really...
[37:40.390 --> 37:40.870]  Naive.
[37:41.190 --> 37:41.670]  Every...
[37:41.670 --> 37:41.930]  Yes.
[37:42.230 --> 37:43.270]  That's very naive.
[37:43.730 --> 37:45.550]  I mean, I think that you would be incredible...
[37:45.550 --> 37:58.490]  As everyone was surprised that Facebook has dozens of people working around the clock to, like, look at the stuff that's coming through the algorithms and sort it all out and, like, make it not totally garbage.
[38:00.410 --> 38:03.890]  Like, all of these companies have staffs of those people.
[38:04.410 --> 38:17.700]  I think that the idea that an algorithm is, like, has no bias because it's computer powered is, like, well, but I mean...
[38:18.250 --> 38:20.370]  But that it's an algorithm that it's...
[38:21.150 --> 38:26.690]  I wouldn't say it's naive because in the Facebook help section, here's the question.
[38:26.690 --> 38:28.830]  How does Facebook determine topics?
[38:29.210 --> 38:30.490]  And they say here...
[38:30.710 --> 38:31.390]  I mean, it's...
[38:31.390 --> 38:36.330]  They don't mention anything about it being an algorithm or it being curated at all, right?
[38:36.390 --> 38:36.590]  Yeah.
[38:36.670 --> 38:39.350]  And most people aren't going to read through this too carefully.
[38:39.350 --> 38:39.470]  Sure.
[38:39.530 --> 38:41.090]  They're trying to focus on their lives.
[38:41.290 --> 38:43.750]  And they're not necessarily diving in deep.
[38:43.750 --> 38:43.950]  Sure.
[38:44.390 --> 38:46.670]  I'm just telling you that, like, I don't...
[38:46.670 --> 39:00.290]  I don't want to live in a world in 2016 where an algorithm, a pure algorithm with no human oversight whatsoever, is controlling everything that comes through my social media feeds because, like, who wrote the algorithm?
[39:00.470 --> 39:00.610]  Yeah.
[39:00.610 --> 39:03.150]  Some squishy human with biases wrote that too.
[39:03.430 --> 39:04.170]  And you know what?
[39:04.250 --> 39:17.990]  There's, like, much less diversity of, like, thought and opinion and, like, background in those algorithm writers than there are amongst the editors and people who are, like, going through and picking out the trends and whatever.
[39:20.410 --> 39:24.030]  And that might be one of the reasons why no one likes to talk about...
[39:25.110 --> 39:25.510]  Yeah.
[39:26.570 --> 39:35.430]  I mean, like, that might be one of the reasons that no one likes to talk about the fact that their technology is not 100% perfect and still churns up regular garbage.
[39:35.730 --> 39:35.910]  Right.
[39:35.910 --> 39:41.690]  If you look at what is the most popular on any social network, I guarantee you it is garbage.
[39:42.810 --> 39:43.330]  Yeah.
[39:43.510 --> 39:44.430]  That's a good point.
[39:44.930 --> 39:46.290]  We need the curators.
[39:46.490 --> 39:46.950]  We do.
[39:47.290 --> 39:47.990]  That's a good point.
[39:49.230 --> 39:49.570]  Hello.
[39:49.750 --> 39:50.310]  First of all...
[39:50.410 --> 40:01.650]  Before I get to my point, I want to commend you for saying that overall people should be able to have different political views without jumping on each other and getting all crazy, whether you're for Trump, Hillary, whoever.
[40:01.930 --> 40:02.650]  Well, thank you.
[40:02.790 --> 40:02.890]  Yeah.
[40:02.890 --> 40:04.670]  I mean, that's just not fair, you know.
[40:05.670 --> 40:06.490]  I'm a registered...
[40:10.030 --> 40:12.470]  And you were very patient with the first guy, by the way, too.
[40:12.510 --> 40:13.350]  I commend you for that.
[40:13.850 --> 40:17.190]  But I'm a registered independent and, you know, I don't follow parties.
[40:17.190 --> 40:23.030]  I kind of go with who I like at the time, you know, concerns the country at the time.
[40:23.550 --> 40:26.430]  But, you know, I sign on AOL.
[40:26.730 --> 40:27.970]  I still have AOL.
[40:28.450 --> 40:30.050]  And there's a giant picture of Hillary.
[40:30.270 --> 40:30.810]  I'm with her.
[40:31.470 --> 40:33.630]  And, you know, I get that's their right to do it.
[40:33.630 --> 40:34.650]  I get on Facebook.
[40:34.650 --> 40:36.410]  And I have, you know, I know my feed.
[40:36.410 --> 40:39.710]  I get my friends and things we're into, movies and stuff.
[40:40.190 --> 40:44.470]  But the sides are inundated with, you know...
[40:44.470 --> 40:45.810]  It doesn't reflect you.
[40:45.930 --> 40:46.870]  It doesn't reflect me.
[40:47.010 --> 40:48.930]  So, obviously, they're not...
[40:48.930 --> 40:50.550]  It's not being catered to me.
[40:50.550 --> 40:56.990]  So, if it's not being catered to me, why is it all stupid things like Laura Ingraham giving a Nazi salute?
[40:57.490 --> 40:59.290]  You know, all Republicans must be racist.
[40:59.710 --> 41:02.970]  Or, you know, even when you get stuff about the Democrats.
[41:03.030 --> 41:05.950]  Oh, Hillary's, you know, jacket cost $6,000.
[41:06.790 --> 41:08.010]  It's just stupidity.
[41:08.290 --> 41:09.050]  And I think it...
[41:09.050 --> 41:09.910]  A fabulous jacket, though?
[41:10.030 --> 41:10.950]  It wasn't fabulous.
[41:11.210 --> 41:11.670]  It was like...
[41:11.670 --> 41:15.150]  It looked like Mao Zedong's jacket back in the day.
[41:15.330 --> 41:17.890]  It's what the curators deemed to be newsworthy.
[41:17.890 --> 41:28.950]  But I look at that, though, and I see, like Facebook, like, you know, if you're an adult who is smart enough to, like, you know, not be affected by what Facebook tells you, what about the youth?
[41:29.170 --> 41:37.650]  You know, I think in general, in the media, music, all facets, there's a certain slant to one party compared to another.
[41:38.970 --> 41:47.930]  And, you know, the slant is kind of fortified with that notion that, well, if you follow Trump, you must, you know, be against gays and you must be against African Americans.
[41:48.410 --> 41:56.710]  And I think that kind of... it kind of, like, subliminally or subconsciously sustains that kind of schism that we have going on now.
[41:56.910 --> 41:57.150]  Yeah.
[41:57.150 --> 41:58.010]  It doesn't do any good.
[41:58.210 --> 42:03.030]  So whether it's going to help Trump or help Hillary in general, you know, it's their right to do it.
[42:03.130 --> 42:05.210]  But it's just, it's troublesome and it's concerning.
[42:05.610 --> 42:08.970]  Well, and maybe it's hurting us in terms of our stereotypes of each other.
[42:09.050 --> 42:10.050]  Oh, I definitely think so.
[42:10.170 --> 42:16.370]  Because I, you know, I can't wait till this election is over just because my friends, it's like everybody's hating each other.
[42:16.630 --> 42:17.610]  Well, thank you for that.
[42:17.950 --> 42:18.610]  Thank you very much.
[42:23.610 --> 42:24.290]  Thanks for your talk.
[42:24.290 --> 42:26.590]  I'm just wondering, you talked a little about the peer-to-peer thing.
[42:26.710 --> 42:28.610]  And I know that was kind of touched on with the filter bubble thing.
[42:28.710 --> 42:34.450]  I'm just wondering if you thought more if there's any, like, way of sort of getting the critical mass for an alternative to...
[42:34.630 --> 42:35.750]  ...a decentralized alternative to Facebook or Twitter.
[42:35.830 --> 42:42.230]  Do you think that's just sort of too difficult so the only thing we can do is to just put pressure on Facebook and Twitter to improve and...
[42:42.230 --> 42:43.670]  I don't want to say it's too difficult.
[42:43.850 --> 42:46.370]  I think that it's absolutely worthwhile to put pressure...
[42:46.370 --> 42:48.450]  Diaspora tried it and isn't doing all that well.
[42:48.610 --> 42:48.810]  Yeah.
[42:49.030 --> 42:52.990]  Well, you know, and I was making social apps for a long time and I stopped.
[42:52.990 --> 43:00.610]  But the reason why I did it was because I thought you could get new companies and new social networks off the ground.
[43:00.790 --> 43:02.390]  And I still believe that we can.
[43:02.810 --> 43:04.950]  It just...it's a really big endeavor.
[43:05.230 --> 43:11.790]  You have to find some way to get so many people on it, right?
[43:11.910 --> 43:19.750]  Like, if you really want it to take off and sustain as a platform, it has to be appealing to a wide group of people.
[43:20.010 --> 43:21.470]  And that's tough.
[43:22.090 --> 43:23.710]  But I think that we may get there.
[43:24.050 --> 43:24.810]  Google Plus?
[43:24.990 --> 43:25.030]  No.
[43:25.290 --> 43:25.710]  Oh my goodness.
[43:25.970 --> 43:26.750]  What's your real name?
[43:26.750 --> 43:28.330]  I think that's the challenge, right?
[43:28.450 --> 43:36.710]  Because so many people here at HOPE have the technical ability and those platforms and the open-source code to build them.
[43:36.750 --> 43:37.890]  That's already out there.
[43:37.890 --> 43:45.010]  The tricky piece is like, how do we make it stick so we can have something that's more ours?
[43:45.450 --> 43:47.190]  And it's a really hard problem.
[43:47.190 --> 43:49.350]  But I do think that we could get there.
[43:49.770 --> 43:50.370]  You know?
[43:50.530 --> 43:51.330]  I really do.
[43:51.630 --> 43:53.310]  And things are changing very fast.
[43:53.450 --> 43:54.030]  Look at the world.
[43:54.210 --> 43:56.050]  Look at the things that you never thought would have happened.
[43:56.050 --> 43:58.150]  We can for sure make a new social network.
[43:58.330 --> 44:00.110]  I'm not sure how just yet.
[44:00.830 --> 44:07.310]  But I think that's absolutely a worthwhile topic for us to keep noodling on.
[44:07.630 --> 44:09.350]  And I have to believe it's possible.
[44:09.950 --> 44:10.270]  You know?
[44:10.350 --> 44:12.070]  I like to think that the future's going to be better.
[44:12.090 --> 44:17.470]  Do you think we can build another one that doesn't just become another corporate Facebook and Twitter?
[44:17.470 --> 44:18.170]  Oh, yeah.
[44:18.510 --> 44:21.850]  I mean like building a decentralized, building like another corporate one is...
[44:21.850 --> 44:22.710]  Oh, but you can.
[44:22.950 --> 44:23.030]  Yeah.
[44:23.270 --> 44:25.730]  Because the thing is, you look at Google, right?
[44:26.190 --> 44:28.210]  This is actually really important.
[44:29.170 --> 44:29.350]  Sorry.
[44:29.630 --> 44:30.170]  Diaspora.
[44:30.510 --> 44:34.570]  If so, you remember the Kickstarter maybe of like...
[44:34.570 --> 44:35.870]  It was a while ago.
[44:36.070 --> 44:41.250]  But Diaspora, they built a decentralized social networking platform.
[44:41.370 --> 44:43.450]  I mean they built it, but did anybody come?
[44:43.890 --> 44:44.170]  No.
[44:44.790 --> 44:47.150]  It's used, but there are people on it.
[44:47.150 --> 44:48.270]  The guy killed himself.
[44:48.510 --> 44:49.010]  I mean...
[44:49.010 --> 44:52.210]  Well, I would argue this about what we could do.
[44:52.350 --> 44:57.350]  If you look at Google, Google became evil, you could argue.
[44:57.610 --> 45:00.470]  It certainly created problems that it didn't anticipate.
[45:00.850 --> 45:08.770]  But Google solved many of the problems it wanted to solve in the early days in terms of like what a work environment could be like.
[45:09.730 --> 45:12.270]  Google did so many wonderful things.
[45:12.570 --> 45:16.190]  They just, they didn't foresee the dark side.
[45:16.190 --> 45:17.630]  And I think that's what we can do.
[45:17.770 --> 45:21.950]  I think we can solve this and we're going to mess up in some way that we won't foresee.
[45:22.310 --> 45:25.930]  But it's still like a process of evolution and things get better.
[45:26.330 --> 45:27.950]  So like I'm down for that.
[45:28.170 --> 45:34.250]  I'm down for us to solve this problem, unintentionally create something that we can't foresee because it didn't exist yet.
[45:34.250 --> 45:36.150]  And then it keeps going.
[45:39.490 --> 45:39.970]  Please.
[45:41.390 --> 45:47.730]  Google does like also like other things that they do that's really good is Google Summer Code which gets college students.
[45:47.970 --> 45:48.290]  Yeah.
[45:48.450 --> 45:50.810]  And then they have Google Code in which actually gets...
[45:50.810 --> 45:51.750]  They do a lot of good.
[45:51.750 --> 45:56.530]  It reaches out to middle school to high school students so that they can get in.
[45:57.550 --> 46:03.630]  So if you actually want to change the way Google and Facebook do their algorithms, I have a very simple solution.
[46:03.970 --> 46:04.370]  You know?
[46:04.510 --> 46:05.250]  Go work for them.
[46:06.250 --> 46:07.050]  They're hiring.
[46:07.270 --> 46:07.910]  Both of them.
[46:08.450 --> 46:09.730]  As does Twitter, I think.
[46:09.930 --> 46:10.530]  Although they kept...
[46:15.120 --> 46:17.300]  No, no, no, no, no, no, no.
[46:17.380 --> 46:18.300]  That is...
[46:19.400 --> 46:21.420]  You obviously were not on the debate team.
[46:21.420 --> 46:22.760]  That was an idiotic remark.
[46:23.700 --> 46:26.220]  If you want to counter it, you're welcome to come up to the mic.
[46:29.220 --> 46:30.520]  So I challenge...
[46:31.120 --> 46:42.660]  I want to challenge the original assumption that Facebook even has a responsibility or we even have the right to demand of them to be neutral.
[46:43.020 --> 46:43.400]  You know?
[46:43.500 --> 46:46.760]  You look at things that are supposed to be neutral because they're...
[46:46.760 --> 46:53.700]  They claim to be generalist organizations like Fox News who are anything bad and all they do is promote hatred.
[46:54.580 --> 46:57.040]  And nobody's actually taking them up to the task.
[46:57.780 --> 47:02.760]  So why should somebody who doesn't even have the...
[47:05.200 --> 47:09.320]  I would be more comfortable if Facebook admitted any bias that it has.
[47:09.460 --> 47:10.480]  That's my point of view.
[47:10.480 --> 47:11.800]  I'm more comfortable in bias.
[47:11.800 --> 47:14.500]  Why is your comfort even a consider...
[47:14.500 --> 47:17.560]  Why should your comfort be a consideration for Facebook?
[47:17.760 --> 47:21.640]  If you don't like Facebook, do what the previous speaker said.
[47:21.880 --> 47:23.020]  Don't be on Facebook.
[47:23.280 --> 47:24.680]  And, you know, you can...
[47:24.680 --> 47:26.240]  There is life outside of Facebook.
[47:26.600 --> 47:27.220]  Fair enough.
[47:27.560 --> 47:28.400]  Next question.
[47:30.620 --> 47:31.580]  So when...
[47:31.580 --> 47:34.480]  I've been thinking about this a little bit and hearing about something kerfuffle.
[47:34.660 --> 47:39.000]  And it seemed like a lot of people that were talking about it kind of weren't really techy.
[47:39.000 --> 47:43.720]  They were saying things like, you can't change the algorithm at all, you know.
[47:44.040 --> 47:44.060]  Ever.
[47:44.340 --> 47:44.560]  Yeah.
[47:44.740 --> 47:48.120]  And it was very much like, you know, whatever Facebook did last Tuesday, well, that was perfect.
[47:49.460 --> 47:54.600]  And that kind of didn't seem to go along with how we have tech companies now that, you know, they're doing A-B tests.
[47:54.720 --> 47:56.640]  They're changing little things all the time.
[47:57.120 --> 47:59.360]  And Facebook is running tests all the time.
[47:59.540 --> 47:59.600]  Yeah.
[47:59.800 --> 48:03.720]  And every, you know, every, you know, Twitter is in every big, you know, tech companies.
[48:03.860 --> 48:04.480]  That's how they work.
[48:05.020 --> 48:05.900]  So what...
[48:05.900 --> 48:07.820]  And I was thinking about this and I'm almost not sure.
[48:08.080 --> 48:09.260]  What do you see...
[48:09.260 --> 48:11.860]  What would a more transparent Facebook be?
[48:12.000 --> 48:17.240]  How would they be able to communicate that, you know, easily to people that are techy and non-techy?
[48:17.920 --> 48:23.920]  Well, if they remove the curated section, that would remove a lot of this problem off the bat.
[48:25.460 --> 48:32.960]  Because I think the way that they handle the algorithm for your feed is generally more transparent and less problematic.
[48:33.080 --> 48:38.300]  People don't seem concerned about what's showing up in their feed outside of search engine optimization.
[48:39.080 --> 48:43.420]  For example, if you're a nonprofit brand, it's hard for you to show up in the feed.
[48:43.580 --> 48:45.740]  But that's not really politicized.
[48:46.400 --> 48:50.220]  And you can choose what shows up in your feed.
[48:50.220 --> 48:55.940]  So for me, like, I have my top 10 or 20 friends and I've chosen that they should get shown first.
[48:55.940 --> 48:57.380]  And they do get shown first.
[48:57.800 --> 49:00.600]  And all in all, that seems to work well enough.
[49:00.820 --> 49:03.580]  Not perfect, but well enough.
[49:04.040 --> 49:11.580]  It's these moments where there's curation against a claim of neutrality that I'm seeing.
[49:13.020 --> 49:13.920]  It's just BS.
[49:14.900 --> 49:15.440]  Thank you.
[49:17.960 --> 49:18.360]  Hello.
[49:19.200 --> 49:23.080]  First off, I'd like to thank you for being patient with this somewhat rowdy group of people.
[49:23.720 --> 49:25.120]  I'm a rowdy group of people.
[49:25.240 --> 49:25.880]  Good, good.
[49:26.040 --> 49:26.760]  And you're in good company.
[49:27.640 --> 49:30.500]  So are you looking for these companies to be completely neutral?
[49:30.780 --> 49:34.580]  Or maybe it's okay if they have a stance, but they're just more transparent about what it is?
[49:34.580 --> 49:41.600]  Well, you know, I'm coming out of having worked in more open-source software over the last few years.
[49:41.600 --> 49:42.960]  And I like transparency.
[49:43.440 --> 49:46.800]  I think that transparency is a nice practice.
[49:47.120 --> 49:49.620]  You know, you make mistakes, but you're open about it.
[49:49.700 --> 49:52.360]  And we can have conversations about it as a community.
[49:52.600 --> 49:54.320]  So that would be my preference.
[49:54.320 --> 50:02.360]  My preference would be for companies to just admit that they're imperfect, and it's a process, and solicit feedback.
[50:02.720 --> 50:09.040]  And for us to say, yeah, we'd rather have you curate, and we understand that humans are messy and the curation's not perfect.
[50:09.260 --> 50:15.280]  And for Facebook to say something like, neutrality is a value, but it's like, we get as close as we can.
[50:15.740 --> 50:21.680]  I think something like that would probably... I'd probably like that a lot.
[50:22.860 --> 50:27.480]  It's the conflict between when a company says a thing, and it's just not what they're doing.
[50:27.680 --> 50:28.560]  That's what troubles me.
[50:29.160 --> 50:32.980]  I'm also just troubled by the enormous power that these companies have in elections.
[50:33.260 --> 50:36.100]  And because of the lack of transparency, we just don't know.
[50:36.700 --> 50:42.120]  I think it's unlikely that Facebook or Twitter will swing this election, or certainly that they'll try.
[50:42.800 --> 50:45.220]  But it remains true that they could.
[50:45.860 --> 50:49.820]  And I think that that's worth... it's worth our attention.
[50:50.920 --> 50:52.280]  Did I answer your question?
[50:52.500 --> 50:52.640]  Yeah, yeah.
[50:52.920 --> 50:55.860]  Also, I have a follow-up, which is, you seem to be saying neutrality is a good value.
[50:56.260 --> 50:57.980]  Is there some limit to neutrality?
[50:58.340 --> 51:04.900]  Like, we can agree that if they choose Democrats or Republicans, that's a bummer, but should they be censoring ISIS posts?
[51:05.140 --> 51:13.720]  Like, is there anything that's so objectionable that it is ethically desirable for them to step in and say, no, we have an opinion on this content or this viewpoint?
[51:13.720 --> 51:19.060]  When I take a step back and I think about Facebook and I think about Twitter, I'm seeing things that I've selected to see.
[51:19.660 --> 51:23.340]  So I shouldn't see ISIS unless I've selected to see them.
[51:23.520 --> 51:26.680]  And then we get into the question of whether Facebook should be hosting that content.
[51:27.200 --> 51:32.540]  And that's something that I'm unsure about.
[51:33.420 --> 51:37.720]  That's a free speech question that I'm not at ease with yet myself.
[51:38.080 --> 51:40.560]  And I'm interested in what other people have to say.
[51:41.720 --> 51:49.100]  If I ran one of those companies, I wouldn't want to host certain kinds of speech, but it is difficult to figure out where you draw certain lines.
[51:49.360 --> 51:51.400]  I suppose ISIS might be an easy line to draw.
[51:51.760 --> 51:53.640]  Other lines are less easy.
[51:55.520 --> 51:57.000]  Yeah, these are good questions.
[51:57.260 --> 51:57.700]  Yeah.
[51:58.460 --> 51:58.600]  All right.
[51:58.680 --> 51:58.840]  Thank you.
[51:59.000 --> 51:59.460]  Thank you.
[52:01.160 --> 52:07.660]  This topic seems to attract a lot of mansplaining to add on to that.
[52:08.100 --> 52:09.340]  Are you saying I was mansplaining?
[52:10.500 --> 52:12.480]  I wasn't calling out anyone in particular.
[52:13.560 --> 52:16.640]  Well, maybe my whole talk was mansplaining and I apologize for that.
[52:17.100 --> 52:17.200]  No.
[52:18.700 --> 52:25.940]  So, my question was to add to someone who asked earlier about how algorithms can have bias.
[52:26.120 --> 52:32.400]  And I wanted to add on to that and say, like, algorithms are, you know, often like simulations.
[52:32.620 --> 52:33.500]  So, if you think about...
[52:33.500 --> 52:43.160]  I was reading an article about the algorithms used to predict whether someone will wind up in jail again after being released.
[52:43.700 --> 52:57.020]  And it was really interesting to see how, like, the inputs into an algorithm really affect the outcome and can actually cause, like, racism or bias in one way or another depending on how the algorithm is crafted.
[52:57.200 --> 53:14.260]  So, do you have any ideas about how we can ask for transparency, especially from, like, government institutions, about the, like, machine learning algorithms or the inputs that they use into those algorithms to solve these problems that are really complicated and,
[53:14.300 --> 53:16.220]  you know, they're, you know, a machine.
[53:16.220 --> 53:19.700]  So, how can a machine have bias but the inputs into those systems?
[53:20.000 --> 53:20.320]  Yeah.
[53:20.660 --> 53:31.410]  So, as the Teeks, who I've been working with on press stuff, is giving a talk tomorrow that deals with how they made changes in Oakland around anti-surveillance and...
[53:31.840 --> 53:32.160]  Sunday.
[53:32.660 --> 53:33.320]  Oh, hi.
[53:33.600 --> 53:33.820]  Yeah.
[53:34.140 --> 53:34.980]  So, Sunday.
[53:35.260 --> 53:35.840]  My apologies.
[53:35.980 --> 53:36.320]  Not tomorrow.
[53:36.440 --> 53:36.640]  Sunday.
[53:36.860 --> 53:43.160]  So, I recommend going and seeing that talk to see how citizens can impact that sort of thing.
[53:43.460 --> 53:49.760]  What I've seen specifically in open-source and transparency is the election of certain candidates who promote those values.
[53:49.920 --> 53:50.920]  And that's starting to happen.
[53:51.280 --> 53:54.660]  Ben Kalos in Manhattan is one example.
[53:55.580 --> 53:58.880]  Rand Paul was interested in open-source when I was working with him.
[53:59.400 --> 54:04.860]  So, I think we can vote for those candidates and make it clear that we care.
[54:05.180 --> 54:08.580]  And perhaps maybe, like, become a voting block as hackers or something.
[54:08.740 --> 54:14.120]  Like, I think that there might be political ways that we can communicate to the reps that we care.
[54:14.480 --> 54:18.360]  And I think we can become louder on specific issues.
[54:18.540 --> 54:25.500]  Like, if people are getting more jail time because an algorithm determines that, that seems like something that citizens should be able to see.
[54:26.060 --> 54:26.640]  All right.
[54:26.860 --> 54:26.960]  Yeah.
[54:27.100 --> 54:27.300]  Thanks.
[54:28.880 --> 54:30.620]  I think we have time for one more question.
[54:33.360 --> 54:34.120]  No one comes up.
[54:34.240 --> 54:34.460]  All right.
[54:34.760 --> 54:35.580]  Thank you so much.