Discover more from Last Week in AI

Weekly text and audio summaries of the most interesting AI news, as well as editorials commenting on recent events

Over 20,000 subscribers

Maine's facial recognition law, hate speech detection on Tiktok, and more!

Last Week in AI #124

Last Week in AI

Jul 12, 2021

Faces in a crowd are overlaid with computer data from facial recognition technology. — Image Credits: David McNew/Getty Images via Slate

Mini Briefs

Maine Now Has the Toughest Facial Recognition Restrictions in the U.S.

As debate over the merits and risks of facial recognition technology continues, Maine's House and Senate unanimously approved LD 1585 on June 16 and 17. LD 1585 is only the second statewide law to regulate facial recognition, but stands out as the nation's toughest law on the matter. Championed by the ACLU, Maine's legislation takes a key step beyond its predecessors by prohibiting use of facial recognition across all levels of state, county, and municipal government. The law outlines strict parameters for exceptions made for law enforcement purposes, and stipulates that facial recognition data cannot be used alone to establish probable cause for an arrest.

How TikTok's hate speech detection tool set off a debate about racial bias on the app

Just as Facebook and Twitter have made efforts to wrangle hate speech on their platforms, so has the upstart social platform Tiktok. But Tiktok's hate speech detector seems to suffer from an error: when Tiktok users typed phrases about Black content in their bios, such as "Black success" or "Black Lives Matter," Tiktok flagged those users' content as "inappropriate." If those users replaced those phrases with "white supremacy" or "white success," they received no warning. As a result, many Black creators have been protesting Tiktok because they feel unrecognized and underserved by the platform, falling into a larger trend of how their content is moderated on social media. A Tiktok spokesperson said the platform is actively working to resolve this issue.

Podcast

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

If you are not subscribed yet and like our weekly AI newsletters, feel free to subscribe below!

News

Advances & Business

How AI is serving up aces at Wimbledon - and what the technology means for the future of sport - "What’s the best way to watch Wimbledon? With more than 18 courts playing up to four matches a day, keeping track of tennis results and highlights can be tricky."
Graphcore's AI chips may not be as powerful as Nvidia's GPUs, but may provide good bang for your buck - "The latest results by benchmarking consortium MLPerf, tracking the best chips for training the most popular neural networks, are out and a new player has entered the game: Graphcore. Each MLPerf release is pretty standard."
Google AI Introduces A Machine Learning Based System For Game Developers To Quickly And Efficiently Train Game-Testing Agents - "Google AI recently announced a machine learning-based framework that game developers could use to train game-testing agents quickly and efficiently, freeing human testers to focus on more complicated problems."
TikTok's Algorithm and AI Tech Are Now Up for Sale - "Now anyone can tap into the secret sauce behind ByteDance’s globally successful TikTok app—for a price."
Facebook & UC Berkeley Substitute a Convolutional Stem to Dramatically Boost Vision Transformers’ Optimization Stability - "Recent studies have shown that vision transformer (ViT) models can attain better results than most state-of-the-art convolutional neural networks (CNNs) across various image recognition tasks, and can do so while using considerably fewer computational resources."
Welcome to Simulation City, the virtual world where Waymo tests its autonomous vehicles - "A light gray cube with a thin blue top glides down a darkened highway, beset on all sides by dozens of green cubes. The green cubes bounce between lanes in an attempt to pass the gray cube, but the gray cube maintains a steady speed as the blackened landscape slips past into the artificial night."
Google AI: Machine Learning for Rapid Training to Game-Playing Agents - "Recently, Google AI has announced that the use of a machine learning system for rapid training to game-playing agents. This framework can be used by game developers to release human game-playing agents, who can efficiently focus on other priority duties to boost productivity."
A Global Smart-City Competition Highlights China’s Rise in AI - "Four years ago, organizers created the international AI City Challenge to spur the development of artificial intelligence for real-world scenarios like counting cars traveling through intersections or spotting accidents on freeways."
Could AI Keep People ‘Alive’ After Death? - "Experts are exploring ways artificial intelligence might confer a kind of digital immortality, preserving the personalities of the departed in virtual form and then allowing them to evolve Resume Subscription"
AnyVision raises $235 million Series C led by SoftBank - "Israeli AI-based face recognition startup AnyVision has completed a $235 million Series C funding round led by SoftBank and Eldridge Industries, Calcalist has learned."
Abundant Robotics shuts down fruit harvesting business - "Abundant Robotics, a Hayward, Calif.-based agricultural robotics company founded in 2016, has shut down. According to a memo about Abundant’s liquidation obtained by The Robot Report, the company “was unable to develop the market traction necessary to support its business during the pandemic."
Pastbook launches a Google Photos-esque AI-powered photobook app - "Tristan covers human-centric artificial intelligence advances, politics, queer stuff, cannabis, and gaming."
What A Long, Strange Trip It's Been: EleutherAI One Year Retrospective - "It’s been a weird year. That’s hardly a rare sentiment for obvious reasons, but it was definitely an especially weird year for a small group of AI enthusiasts on a niche discord server. It’s been one year since the founding of EleutherAI (back then still called LibreAI)."
AI Text-to-Speech models have accents too - "While playing around with the Web Speech API I discovered something interesting: if you let the Dutch voice speak English text, it sounds remarkably like the typical "Dungrish" accent (Dutch person speaking English). I didn't expect this."

Concerns & Hype

Narratives of AI behind the Iron Curtain — AI Narratives - "Jędrzej Niklas, a research associate at the Data Justice Lab at Cardiff University, concluded the workshop with a presentation on the role of data-driven technologies in communist Poland and their perceived utopian potential to govern the state."
Google AI Introduces A Dataset for Studying Gender Bias in Machine Translation - "Neural machine translation (NMT) has been growing with leaps and bounds, and it has also enabled natural and fluid translations to a considerable degree."
Tencent Deploys Facial Recognition to Detect Minors Gaming at Night - "China’s gaming giant Tencent Games has launched one of the most-talked about — and controversial — means to curb excessive play time among children and teenagers so far: a time-sensitive facial recognition system."
This AI publicly shames politicians, but don’t laugh just yet - "A Belgian artist has devised a novel way to catch misbehaving politicians: automatically tagging the rogues who use their phones at work."
Google AI chief Jeff Dean sparks cries of hypocrisy as he urges marginalized groups to work with its researchers: 'After what you did to Timnit?' - "Google's lead exec for artificial intelligence sparked an online backlash after encouraging students from historically marginalized groups to apply to a mentorship program at the company, despite an ongoing feud with his team's former ethics leads."
The Criminal Potential of Artificial Intelligence - "AI can be implicated in crime in several ways."
YouTube’s recommender AI still a horrorshow, finds major crowdsourced study - "For years YouTube’s video-recommending algorithm has stood accused of fuelling a grab-bag of societal ills by feeding users an AI-amplified diet of hate speech, political extremism and/or conspiracy junk/disinformation for the profiteering motive of trying to keep billions of eyeballs stuck to its"
Attackers use ‘offensive AI’ to create deepfakes for phishing campaigns - "Where does your enterprise stand on the AI adoption curve? Take our AI survey to find out. AI enables organizations to automate tasks, extract information, and create media nearly indistinguishable from the real thing. But like any technology, AI isn’t always leveraged for good."
The first WHO report on AI in healthcare is a mixed bag of horror and delight - "The World Health Organization today issued its first-ever report on the use of artificial intelligence in healthcare."
Self-Driving Teams Have Always Strived To Measure Safety. What If That’s Not The Hard Thing? - "Since the first dream of self-driving, there has always been a sharp focus (including from myself) on safety. The ability to reduce accidents is the grandest purpose of these projects, and every team declares that safety is at the core of everything they do."
OpenAI warns AI behind GitHub’s Copilot may be susceptible to bias - "Where does your enterprise stand on the AI adoption curve? Take our AI survey to find out. Last month, GitHub and OpenAI launched Copilot, a service that provides suggestions for whole lines of code inside development environments like Microsoft Visual Studio."
Data labeling for AI research is highly inconsistent, study finds - "Where does your enterprise stand on the AI adoption curve? Take our AI survey to find out. Supervised machine learning, in which machine learning models learn from labeled training data, is only as good as the quality of that data."

Analysis & Policy

Europe makes the case to ban biometric surveillance - "Companies are racing to track your emotions, how you walk and your voiceprint. Should Europe ban biometric tracking entirely?"

Expert Opinions & Discussion within the field

Elon Musk just now realizing that self-driving cars are a ‘hard problem’ - "Tesla CEO Elon Musk is finally admitting that he underestimated how difficult it is to develop a safe and reliable self-driving car. To which the entire engineering community rose up as one to say, 'No duh.' Or at least that’s it should have happened in a just world."
America's global leadership in human-centered AI can't come from industry alone - "The Biden administration has followed through on a Congressional mandate to create a National AI Research Resource Task Force."