Mini Briefs

Controversial facial-recognition software used 30,000 times by LAPD in last decade, records show

It is now well-known that a number of police departments around America have made use of facial recognition. But the Los Angeles Police Department’s has been a particularly heavy user: the LAPD used facial recognition software nearly 30,000 times since 2009. The figures reveal the extensive use behind the vague and contradictory information the LAPD provided about how and whether it uses the technology–the department often denied using facial recognition at all. While the LAPD does not have its own platform, it does have access to facial recognition software through a regional database. The LAPD now claims it was open about its use, but civil liberties advocates note the LAPD’s recent denials which were only corrected after being questioned.

Twitter and Zoom’s algorithmic bias issues

Social media platform Twitter has come under fire recently for an image cropping algorithm that users noticed automatically focused on white faces over black ones. While Twitter claims it tested the service for bias before deploying it, the company accepts that it didn’t do enough. Video conferencing service Zoom received criticism for similar reasons after a PhD student tweeted about a Black faculty member’s issues with the platform: when using a virtual background, Zoom would remove his head. There has been debate about the causes of these complications and whether Twitter’s cropping bias is real, but the fact that it crops out Black people at all when it does not crop out white people renders its algorithm problematic.

Other Sources:

• Twitter apologises for ‘racist’ image-cropping algorithm
• Why Twitter’s image cropping algorithm appears to have white bias

Podcast

Check out our weekly podcast covering these stories! Website | RSS | iTunes | Spotify | YouTube

News

Advances & Business

• YouTube Removed Twice As Many Videos After Switch to AI Moderation - YouTube removed twice the usual amount of videos between April and June after relying on AI to moderate content amid forced lockdowns. The increase in video removals is a result of YouTube reducing human oversight during the content moderation process.

• At the Math Olympiad, Computers Prepare to Go for the Gold - Computer scientists are trying to build an AI system that can win a gold medal at the world’s premier math competition. The 61st International Mathematical Olympiad, or IMO, begins today.

• Microsoft teams up with OpenAI to exclusively license GPT-3 language model - One of the most gratifying parts of my job at Microsoft is being able to witness and influence the intersection of technological progress and impact: harnessing the big trends in computing that have the opportunity to benefit everybody on the planet.

• AI devs created a lean, mean, GPT-3-beating machine that uses 99.9% fewer parameters - AI researchers from the Ludwig Maximilian University (LMU) of Munich have developed a bite-sized text generator capable of besting OpenAI’s state of the art GPT-3 using only a tiny fraction of its parameters.

• To Make Fairer AI, Physicists Peer Inside Its Black Box - Physicists built the Large Hadron Collider to study the inner workings of the universe.

• Microsoft’s new feature uses AI to make video chat less weird - How Microsoft’s Surface Pro X uses driver-based software and dedicated AI hardware to make video chat more people-friendly. With so many of us working from home, we’ve shifted into a world where video conferencing has become the main way we connect with colleagues.

• AI planners in Minecraft could help machines design better cities - A dozen or so steep-roofed buildings cling to the edges of an open-pit mine. High above them, on top of an enormous rock arch, sits an inaccessible house. Elsewhere, a railway on stilts circles a group of multicolored tower blocks. Ornate pagodas decorate a large paved plaza.

• Facebook wants to make AI better by asking people to break it - The new kind of test pits machine-learning models against humans who do their best to fool them. The explosive successes of AI in the last decade or so are typically chalked up to lots of data and lots of computing power.

Concerns & Hype

• People’s notions about AI are terrible, an MIT study asks whether they can be helped - One of the most striking PR moments of the AI age was the sale by Christie’s auction house in October, 2018, of a painting output by an algorithm, titled “Edmond de Belamy,” for $432,000.

• Diversity in AI: The Invisible Men and Women - In June, a crisis erupted in the artificial intelligence world.

• How humane is the UK’s plan to introduce robot companions in care homes? - Some UK care homes are to deploy robots in an attempt to allay loneliness and boost mental health. The wheeled machines will “initiate rudimentary conversations, play residents’ favorite music, teach them languages, and offer practical help, including medicine reminders.” They are being introduced after an international trial found they reduced anxiety and loneliness.

• We’re not ready for AI, says the winner of a new $1m AI prize - Regina Barzilay, the first winner of the Squirrel AI Award, on why the pandemic should be a wake-up call.

• A controversial photo editing app slammed for AI-enabled ‘blackface’ feature - Photo editing app Gradient is under fire for a new feature that lets people alter their ethnicity in images, with many slamming it for promoting digital “blackface.

• These weird, unsettling photos show that AI is getting smarter - Models are learning how to generate images from captions, a sign that they’re getting better at understanding our world. Of all the AI models in the world, OpenAI’s GPT-3 has most captured the public’s imagination.

Analysis & Policy

• CSET Publishes AI Policy Recommendations for the Next Administration - As policymaker attention toward global competition in emerging technologies grows, CSET has compiled the following one-pagers with recommendations for addressing critical issues affecting U.S. and overseas development of artificial intelligence.

• Exclusive: New push for autonomous vehicles bill - In a renewed push to get an autonomous vehicles bill through Congress, Rep. Bob Latta (R-Ohio) is reintroducing the SELF Drive Act Wednesday, Latta told Axios. The big picture: New policy legislation is a long shot in the short Congressional calendar leading up to Election Day.

• AI Commission Wants to Know How Government Can Help Industry Boost Commercial Innovation - The National Security Commission on Artificial Intelligence invited small- and medium-sized AI-focused firms to share their thoughts on how the government should work with industry to fortify commercial innovation of the now widely-used, evolving technology.

Expert Opinions & Discussion within the field

• My data point as a woman in AI - I have not felt disadvantaged for being a woman in AI. Two immediate clarifications: This is not a counter data point to the experiences other women in AI have shared. It is simply an additional one. My experience is likely a consequence of obliviousness, luck, and privilege.

That’s all for this week! If you are not subscribed and liked this, feel free to subscribe below!