aiTransparency's profile picture. Our organization informs citizens regarding the workings and consequences of the algorithms that shape our access to information.

AlgoTransparency

@aiTransparency

Our organization informs citizens regarding the workings and consequences of the algorithms that shape our access to information.

Pinned

Now working on AlgoTransparency V2 to provide a better understanding of the impact of YouTube's AI on information consumption In the meantime, here is some recent media coverage of our work: algotransparency.org/en/press.html


AlgoTransparency reposted

Another day. Another promise Facebook isn’t keeping.

Feb. 8: Facebook promises to remove vaccine misinformation. Nov. 15–May 1: We tracked 31 times #CitizenBrowser users were recommended anti-vax or anti-mask groups to join, including 7 after Feb. 8. Read the story now from @darakerr and @corintxt. themarkup.org/citizen-browse…

themarkup.org

Facebook Said It Would Stop Recommending Anti-Vaccine Groups. It Didn’t – The Markup

As COVID-19 surged, the social media company pointed users to all sorts of anti-vaccine and anti-mask groups



AlgoTransparency reposted

"Unless we dismantle the increasingly centralized chokepoints in our global digital infrastructure, we can anticipate an escalating political battle between political factions and nation states to seize control of their powers." eff.org/deeplinks/2021…


AlgoTransparency reposted

BREAKING: @TWITTER just published the full dataset of tweets from the Russian and Iranian troll farms. 9 million Russian tweets. 1 million Iranian tweets. #motherlode They gave @DFRLab a look ahead of time. Thread on the findings to follow.


AlgoTransparency reposted

THIS JUST IN: How did misinformation spread during the 2016 presidential election and has anything changed since? One of the largest studies yet on the topic, supported by Knight Foundation and produced by Matthew Hindman of GWU in collaboration with Vlad Barash of #Graphika.

FACT: The vast majority of Twitter accounts that repeatedly spread “fake news” during the 2016 election campaign are still active today, and they publish more than a million tweets / day—via kf.org/misinfo, @MattHindman @Graphika_Inc

knightfdn's tweet image. FACT: The vast majority of Twitter accounts that repeatedly spread “fake news” during the 2016 election campaign are still active today, and they publish more than a million tweets / day—via kf.org/misinfo, @MattHindman @Graphika_Inc


The Markup is one of the most exciting things coming up this year, and will have tremendous positive impact by detecting biases in algorithms.

.@NellieBowles has a very nice write-up in the @Nytimes on our plans to build a newsroom around the scientific method, how we aim to measure and quantify the human impacts of our tech-obsessed society and how I think tech compares to canned food! nytimes.com/2018/09/23/bus…



AlgoTransparency reposted

My favorite example of how informationally toxic YouTube's algorithm is this: Imagine you're high school freshman and got a school assignment about the Federal Reserve.


AlgoTransparency reposted

Really important work here from @gchaslot. The fact that AI optimized for watch time promotes content that says the media is lying should terrify you. In other words, Youtube, FB, Twitter, are all more likely to push this kind of messaging.

algorithms that try to maximize how much time you spend on their platform will incentivize content that says the rest of the media is lying @gchaslot medium.com/@guillaumechas…

math_rachel's tweet image. algorithms that try to maximize how much time you spend on their platform will incentivize content that says the rest of the media is lying
@gchaslot medium.com/@guillaumechas…
math_rachel's tweet image. algorithms that try to maximize how much time you spend on their platform will incentivize content that says the rest of the media is lying
@gchaslot medium.com/@guillaumechas…


AlgoTransparency reposted

Relying on algorithms alone will not work. Not today, not tomorrow. And we know providing broader reach isn’t enough. We need to figure out how to help with economic incentives too. We’re behind on that, but thinking deeply about it. Open to great and scalable ideas.


AlgoTransparency reposted

New! MiniMap of @YouTube with 8k channels / 3,600,000,000,000 views. Links between channels are recommendations from YouTube's AI. @YouTube displays only a handful videos at a time. This map shows a broader picture algotransparency.org/youtube.html


AlgoTransparency reposted

BREAKING: In a huge victory, the European Parliament has voted 318-278 against #Article13 and #Article11—the disastrous #CensorshipMachine and #LinkTax copyright proposals. That means we’re close to stopping these terrible proposals—and we’re gaining momentum.


AlgoTransparency reposted

You need to know what's happening in the EU on July 5—it could alter the Internet forever. eff.org/deeplinks/2018…


AlgoTransparency reposted

The sweeping new law passed in CA today could force a massive change in the relationship between consumers and tech companies. Californians can opt out of data collection, have the right to know what data is collected, delete their data, and companies that fail get major fines.


AlgoTransparency reposted

Your cell phone says a lot about where you’ve been. So much so that the Supreme Court said police need a warrant to access that information. eff.org/deeplinks/2018…


A new bias audit tool for machine learning developed by the University of Chicago: dsapp.uchicago.edu/aequitas/


AlgoTransparency reposted

We're proud to announce the commencement of collaboration with Dr Guillaume Chaslot @gchaslot, founder of @HelloAlgo & former @Google @YouTube software engineer, which will initially focus on sharing of data and expertise to fight disinformation. Looking forward to it!

DisinfoEU's tweet image. We're proud to announce the commencement of collaboration with Dr Guillaume Chaslot @gchaslot, founder of @HelloAlgo & former @Google @YouTube software engineer, which will initially focus on sharing of data and expertise to fight disinformation. 
Looking forward to it!

AlgoTransparency reposted

AI+ethics is important, but has been partly hijacked by the AGI (artificial general intelligence) hype. Let's cut out the AGI nonsense and spend more time on the urgent problems: Job loss/stagnant wages, undermining democracy, discrimination/bias, wealth inequality.


.@randhindi’s small experiment to understand privacy

Today marks the coming into effect of #GPDR. @randhindi CEO @snips, tried a little experiment here at #VivaTech, to see how important privacy is to you. Would you be comfortable with this? 📱



AlgoTransparency reposted

I believe none of the most important problems facing the world today are technical problems. All of them are political problems -- we know what the solutions are, we have the technology, and now we need to overcome human greed, fear, and hate, to finally take action.


AlgoTransparency reposted

Real-time location data is some of the most sensitive personal information that wireless carriers have about their customers. They should protect it much better than it seems they are. slate.com/technology/201…


Loading...

Something went wrong.


Something went wrong.