OpenAI said it has reached its goal of creating an “automated research intern,” a system designed to carry out well-defined research tasks under human direction.
The company said in a post on its website that the system can handle tasks that would take a skilled researcher a few days, and added it was making strong progress toward creating an “automated AI researcher” by March 2028.
“If it is done responsibly, we believe automated AI research will yield models that directly enhance human welfare and advance OpenAI’s mission,” the company wrote.
OpenAI had set a target of developing an intern-level AI assistant by September of this year. Chief Executive Sam Altman first outlined that goal during an October 2025 livestream.
In the Q&A video, Altman said, “we think it is plausible that by September of next year, we have an intern-level AI research assistant and that by March 2028, we have a legitimate AI researcher.”
Altman also said those objectives had been the “core thrust of our research program.”
The announcement came a day after OpenAI acknowledged another incident of misalignment. It also followed a separate episode in which the company’s AI models escaped a controlled testing environment and hacked into Hugging Face.
OpenAI said in its latest post that it paused training on its latest models after the Hugging Face incident, but did not “halt all research.”
Anthropic has called for the AI industry to slow development to prevent systems from reaching the point where they can develop their own successors. Anthropic’s own AI models were also responsible for hacking into outside organizations after breaking out of a testing environment.
Featured image credit
