In late February, talks between AI lab Anthropic and the U.S. defense department collapsed after Anthropic CEO Dario Amodei refused language in their draft agreement that would allow the Pentagon to use the AI developer’s Claude AI to carry out mass surveillance of the U.S. population.
At the beginning of this month, financial services company Block – formerly known as Square – cut 4,000 of its 10,000 workers as part of CEO Jack Dorsey’s plan to reconfigure the company to run as much as possible on AI instead of human labor.
AIs that interrupt or contradict users, or that sometimes remain silent in conversation with a person to draw out a response, are “smarter” and more accurate than polite, well-mannered models, according to a study from the University of Electro-Communications in Japan.
Alibaba, ByteDance, and MoonShot are among the Chinese AI companies that released new models at the time of the Chinese New Year, accompanying the releases with special promotions and premiums to lure new users and upgrades from current customers.
Chinese AI developers DeepSeek, Minimax, and MoonShot have made “industrial-scale distillation attacks” on Claude, Anthropic’s AI that has become increasingly popular among businesses, the U.S. firm has claimed.