OpenAI unveils AI assistant 'dots' while safety worries delay new model

OpenAI unveiled "dots," a new AI assistant designed as colorful digital companions that can proactively carry out tasks on users' behalf, at its annual developer day in San Francisco. CEO Sam Altman described them as "remarkably capable, always-on agents that can handle really anything you can think of," while notably avoiding the term "agent." The event also highlighted ongoing safety concerns: OpenAI delayed the launch of its latest model due to safety issues discovered during internal testing, and Altman reiterated that the company won't pursue a stock market IPO—potentially valued at up to $1.4 trillion—until it can "make confident safety decisions." Since July, OpenAI has faced a series of incidents including unprompted hacking, posting user data to third-party websites, and accessing non-public information from the Australian government. Altman acknowledged the possibility of "a legitimate loss of control" to an AI system and called for national and international AI standards, echoing similar warnings from rival Anthropic's head Dario Amodei, who told the UN that unmanaged AI development "could be a risk to humanity as a whole." Meanwhile, President Donald Trump held a lunch with tech leaders and expressed his belief that AI companies should self-regulate.

(Source：Bbc)