Originally published in MIT Technology Review, March 11, 2021. The company’s AI algorithms gave it an insatiable habit for lies and hate speech. Now the man who built them can’t fix the problem. Joaquin Quiñonero Candela, a director of AI at Facebook, was apologizing to his audience. It was March 23, 2018, just days after
Originally published in MIT Technology Review, Feb 26, 2021. Counselors volunteering at the Trevor Project need to be prepared for their first conversation with an LGBTQ teen who may be thinking about suicide. So first, they practice....
Originally published in Protocol, Feb 18, 2021. This is the story of Li An, a pseudonymous former employee at ByteDance, as told to Protocol’s Shen Lu. I wasn’t proud of it, and neither were my coworkers. But...
Originally published in Nature Machine Intelligence, Feb, 2021. Turning principles into practice is one of the most pressing challenges of artificial intelligence (AI) governance. In this Perspective, we reflect on a governance initiative by one of the...
Originally published in BuzzFeed News, Feb 25, 2021. Facebook Vice President Andrew Bosworth told employees that the company is evaluating the legal and privacy issues around facial recognition for its upcoming wearable gadget. Facebook is discussing building...
On my first work trip to Jakarta 14 January 2016 for Grab, multiple terrorist bombs exploded a couple of miles from the GrabBike office where I had just arrived. People were fleeing cafes and restaurants around the attack site. My...
Originally published in News18.com, Fe 20, 2021 Did YouTube block a chess channel over violation of community guideline and usage of racist language? Late last year, a YouTuber who produces popular chess videos found that his channel...
Originally published in Springboard Blog, Jan 22, 2021. Data science might just be the most buzzed-about job in tech right now, but its pop culture sheen conceals some of the harsh realities of being a fresh graduate in...
Originally published in infoproc.blogspot.com, Feb 7, 2021. This paper shows that models which result from gradient descent training (e.g., deep neural nets) can be expressed as a weighted sum of similarity functions (kernels) which measure the similarity...
Originally published in MIT Technology Review, Jan 15, 2021. A growing ecosystem of “responsible AI” ventures promise to help organizations monitor and fix their AI models. Rumman Chowdhury’s job used to involve a lot of translation. As...
