Mini Briefs

Facebook’s AI for detecting hate speech is facing its biggest challenge yet

Facebook has made significant progress recently to proactively take down content that violate its community standards. For example, in the second quarter of 2020, Facebook took down 104.6 million pieces of content. While reviews are typically performed by a vast workforce of human moderators, AI-powered tools have enabled Facebook to do this work at a greater scale for textual content.

However, there’s a long way to go for these systems to match or exceed the capabilities of human moderators. This is because a large proportion of hate speech and misinformation is in the form of images and memes, and reasoning about the context and language-image interplay is an extremely difficult challenge for AI.

Given Facebook’s scale and the speed at which some use it to spread hate, incite violence, and share lies with millions, Facebook will have to keep running to catch up.

AI Slays Top F-16 Pilot In DARPA Dogfight Simulation

The Defense Advanced Research Project Agency (DARPA) recently hosted a simulated F16 dogfight competition, with different AI bots competing with each other as well as with human pilots. The top AI bot was able to beat a human pilot 5-0 in the simulated contest. DARPA started this program “as a risk-reduction effort […] to flesh out how human and machine pilots share operational control of a fighter jet to maximize its chances of mission success.” Competition runners are broadly optimistic about the demonstration of AI capabilities, even if they are not close to being deployed on a real aircraft. Of concern, the program had little discussion on the ethics of AI military applications, especially with the lethal autonomous weapon systems being considered.

Podcast

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

News

Advances & Business

• Microsoft, Energy Dept. to Develop Disaster-Response AI Tools - The U.S. Department of Energy and Microsoft Corp. on Tuesday announced a partnership to develop artificial-intelligence tools aimed at helping first-responders better react to fast-changing natural events, such as floods and wildfires.

• Coronavirus: Robot CERi is a bilingual Covid-19 expert - Ceri is bilingual, clued-up on coronavirus and can tell what mood you are in. Ceri also happens to be a robot.

• Moscow DOH uses AI platform to detect lung cancer symptoms - Moscow’s department of health is using an artificial intelligence (AI) platform to detect symptoms of lung cancer in CT scans, as part of a project to implement AI technology for radiology.

• Scientists develop artificial intelligence system for high precision recognition of hand gestures - The recognition of human hand gestures by AI systems has been a valuable development over the last decade and has been adopted in high-precision surgical robots, health monitoring equipment and in gaming systems.

• Forget credit cards - now you can pay with your face. Creepy or cool? - A new way to pay has arrived in Los Angeles: your face.

• Singular - A collection of stories about AI and singularity co-written by a human and GPT-3

Concerns & Hype

• The dystopian tech that companies are selling to help schools reopen sooner - This fall, AI could be watching students social distance and checking their masks. Thousands of schools nationwide will not be reopening this fall.

• NYPD Used Facial Recognition Technology In Siege Of Black Lives Matter Activist’s Apartment - The NYPD deployed facial recognition technology in its hunt for a prominent Black Lives Matter activist, whose home was besieged by dozens of officers and police dogs last week, a spokesperson confirmed to Gothamist.

• Machines can spot mental health issues - if you hand over your personal data - Digital diagnosis could transform psychiatry by mining your most intimate data for clues. But is the privacy cost worth it?

• Supporting Black Artists Who Are Examining AI - Technology has a complicated relationship with racial justice. Smartphones, internet platforms, and other digital tools can be used to document and expose racism. But digital tools can also fuel racism: smart doorbells surveil Black individuals.

• A-level and GCSE results in England to be based on teacher assessments in U-turn - All A-level and GCSE results in England will be based on grades assesed by teachers instead of algorithms.

Analysis & Policy

• GPT-3 and The Question of Automation - Automation is not an all or nothing proposition. An AI model’s automation capability is highly conjoined with the task and application it is used in.

• An A.I. Movie Service Could One Day Serve You a New Custom Film Every Time - How long will it be until an A.I. can make an actual feature film on demand?

• Fairness, evidence, and predictive equality - How the causal fairness principle relates to predictive equality

• How robotics and automation could create new jobs in the new normal - Depending on who you ask, AI and automation will either destroy jobs or create new ones. In reality, a greater push toward automation will probably both kill and create jobs - human workers will become redundant in certain spheres, sure, but many new roles will likely crop up.

Expert Opinions & Discussion within the field

• Too many AI researchers think real-world problems are not relevant - The community’s hyperfocus on novel methods ignores what’s really important.

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