Happy Sunday and welcome to InsideAI! I'm Rob May, a partner at PJC, focused on AI and Robotics. I write the Sunday commentary for the newsletter here, and longer commentary twice a month at Technically Sentient. This week in the commentary section I am focused on the current "AI Lull".
But before we get to that, let's look at the most clicked articles from this week's newsletters:
Facebook launched Dynabench, a new way to benchmark AI systems by gauging how easy the system is fooled by humans. The Dynabench team argues that the need for new testing arose because existing benchmarks don't actually capture AI performance very well.
MIT professor Regina Barzilay won the $1m Squirrel AI Award for her work on machine learning models to develops drugs and diagnose breast cancer. Barzilay, a breast cancer survivor, is a professor of electrical engineering and computer science at MIT’s Computer Science and Artificial Intelligence Laboratory.
The U.S. Department of Homeland Security (DHS) failed to safeguard data during a facial recognition technology pilot, which led to a 2019 breach that compromised 184k travelers' images, according to a government report. The hack of facial recognition data from Customs and Border Protection (CBP) subcontractor Perceptics resulted in at least 19 other those images being posted on the dark web.