One of my favourite traditions at Meesho is user-calling. Almost every new joinee is expected to call ~50 users before they start working on any project. It's crazy how many 'fancy' ideas go to trash within 10 minutes into the call with a Tier-3 user. ⚡️

Very grateful to Ifeoma Ozoma @erikashimizu @timnitGebru (@dair-community.social/bsky.social) @chelseyglasson and @_jack_poulson for talking to me about what their lives have been like since coming forward with their stories. protocol.com/big-tech-whist…

Close. AI has plenty of doubt (most models can return probabilities for any prediction, if you configure them to). The problem is humans, not doubting the AI enough to notice when it doubts itself. It’s a tool, and it matters how you use it.

Of the 65L vaccine doses recieved, we have already utilised 62,36,676. 2L+ vaccines are being administered daily with almost 0 wastage. Only less than 3L vaccine doses remain with us now. Kerala's demand for 50L doses should be allowed immediately. Ministry of Health

Which law? 🤔

gizmodo.com/signal-tried-t…

This work was not led by NIST. It was led by Dr. Joy Buolamwini, Deb Raji, and @timnitGebru (@dair-community.social/bsky.social). Really disappointed to see this 60 Minutes segment. Joy says she spoke to the producers for many hours and yet they still chose not to feature her work.

Reminder that we are hiring for the ML Ethics team at Twitter! I’m looking for a Staff/Sr. Staff Engineer and an Engineering Manager. Job postings: careers.twitter.com/en/work-for-tw…

Looking forward to this - starting in a few minutes #CogX2021

#Fairness in ML - Sandra Wachter [email protected] and Brent Mittelstadt @bmittelstadt.bsky.social bias preserving and bias transforming #CogX2021

"This is about recognising that we have very bad metrics and proxies for determining merit" Sandra Wachter [email protected] making a case for bias transforming as a way to address indirect discrimination #CogX2021 session on 'How fair are bias detection tests?'

Me, ruchowdh.bsky.social, and Ricardo Baeza-Yates all walk into a bar... What happens next?! Come check it out tomorrow! =) Wonderful to have this opportunity, thanks to @lucyh, BBC Digital Planet, Dr. Joy Buolamwini, @timnitGebru (@dair-community.social/bsky.social), Deb Raji, and everyone else who made this come together.

This is so dubious and tiring (h/t Tim Maughan). While I'm not defending anyone or anything, in using "looking at phone" as a proxy for "distraction" is also misses the reality that texting is how you sidebar with your people during these long often performative sessions.

Scientists: 1970s: We’re seeing concerning patterns. 1980s: Yep, definitely worrisome. 1990s: Seriously folks, this could get bad. 2000s: Do we need to explain this another way? We’re very alarmed. 2010s: *marching in the streets in lab coats* 2020s: bbc.com/news/science-e…

The best thing about switching your web search to ChatGPT is then the people who put all this info on the web don't get any credit and you have no idea how reliable the info is since all sources are jumbled into a giant opaque probability distribution. Oh wait, that's terrible.

My op ed about the last days of Twitter is out. It was sad to reflect on all that could have been. I write this piece not just as a commemoration but as a warning - we need better systems of tech oversight theatlantic.com/technology/arc…

Students in my senior seminar on #AIEthics performed an external audit of OpenAI chatGPT and DALL-E, assessing these services for ethical and social impact across a number of adversarial tests. I’ve compiled some of the results below. 🧵

Using the Ethical Risk Chart from the Smile Detection example case study in Raji et al (2020), the students auditing chatGPT summarized the results of their testing in the chart below.