4 ms·
I Designed Algorithms at Facebook. Here’s How to Regulate Them
- raxxorrax 5y agoTo be honest, people that worked for Facebook perhaps made a mistake in their lives and should be asked last. > As Ms. Haugen said, it should now be humans, not computers, “facilitating who we get to hear from.” Wrong juxtaposition. Ms Haugen argue it should be a government institution that manages this. This is ridiculous and should be denied by any responsible member of the legislative.
- moe44 5y agoEngagement based rankings and recommendations is real mindless garbage. And it is the right thing to reform. Like the mindless stream of randomness HN produces.
- impostervt 5y agoIt seems like it would be much harder to discover other content I'm interested in though, if it were removed.
- prirun 5y ago> The solution is straightforward: Companies that deploy personalized algorithmic amplification should be liable for the content these algorithms promote. I just did a search for the word mob at Google and got 2.5B results. An algorithm has to decide which 50 are going to be displayed first. That may or may not be personalized. For example if I previously search for Al Capone, this search for mob could use that to filter the 2.5B results. If Google does that, does it become liable for all the results it displays? Is it really "promoting" something here, or just trying to satisfy my search with the most relevant results? It seems a very difficult distinction to make. I've never used Facebook and think it's a waste of time. But if they do decide to regulate it, I don't see much difference between the radicalism that gets promoted on Facebook and that on every news program on TV (which I no longer watch), or the radicalism promoted by politicians (which I no longer vote for).
- etblg 5y agoIt does raise a bunch of interesting scenarios that you could potentially sue over. If a Jewish person gets recommended Mein Kampf on Amazon, could they now sue over the hateful content they're being shown? Should Amazon preemptively never show anyone content they could potentially be offended by (i.e. a Jewish person would never be able to be recommended Mein Kampf, a transgender person would never be able to see Ace Ventura under "watch next" on Amazon Prime while watching comedy movies, people in Winnipeg will never have Venetian Snares' "Winnipeg Is A Frozen Shithole" show up on Spotify radio)
- etblg 5y agoThe solution, pretty far down in the story: > The solution is straightforward: Companies that deploy personalized algorithmic amplification should be liable for the content these algorithms promote. This can be done through a narrow change to Section 230, the 1996 law that lets social media companies host user-generated content without fear of lawsuits for libelous speech and illegal content posted by those users. So basically just ban an algorithmic feed by opening it up to endless lawsuits: > There is no A.I. system that could identify every possible instance of illegal content. Faced with potential liability for every amplified post, these companies would most likely be forced to scrap algorithmic feeds altogether. > Social media companies can be successful and profitable under such a regime. Twitter adopted an algorithmic feed only in 2015. Facebook grew significantly in its first two years, when it hosted user profiles without a personalized News Feed. Both platforms already offer nonalgorithmic, chronological versions of their content feeds. Notably Twitter was (AFAICT) losing money in 2015 and only started turning a profit 2-3 years later. Facebook grew in its first two years when it was a new hip young social network aimed at an exclusive group. > This solution would also address concerns over political bias and free speech. Social media feeds would be free of the unavoidable biases that A.I.-based systems often introduce. Any algorithmic ranking of user-generated content could be limited to nonpersonalized features like “most popular” lists or simply be customized for particular geographies or languages. Fringe content would again be banished to the fringe, leading to fewer user complaints and putting less pressure on platforms to call balls and strikes on the speech of their users. The top lists would no longer be personally customizable, but would still offer the same incentives to get to the top that would drive outrage-engagement anyway. Presumably the content that gets boosted would include Fox News (the top rated cable news network), Trump's twitter (when it existed), at least in Georgia Marjorie Taylor Greene (the sitting congresswomen that is an QAnon aficionado), etc. It's trading one algorithm that rewards engagement with another dumber more general one, that I believe would just be mostly the same content anyway. During the Rohingya genocide, the state armed forces were the ones committing genocide on behalf of the majority population, wouldn't the simple nonpersonalized "most popular" list for that geographic region just end up being the same thing?