Sam Altman Announces Internal Goal: AI Capable of Independent Research by March 2028

According to monitoring by 1M AI News, OpenAI CEO Sam Altman confirmed during a company forum that OpenAI has set March 2028 as the official target for achieving an ‘automated researcher,’ an AI system capable of conducting independent AI research. He stated that once realized, the impact would be twofold: the system itself would demonstrate that AI is capable of high-level cognitive work, and it would also accelerate the pace of subsequent AI research, with ‘progress in the following year almost certainly faster than our current speed.’ Altman compared the current moment to the eve of COVID in early 2020. He recalled that by the end of January 2020, OpenAI researchers were more aware than the outside world that the pandemic would change everything, working from home early and disinfecting door handles with copper, which was mocked by the media. One night, he walked alone in San Francisco’s Mission District, watching people in bars breathe face-to-face through the windows, realizing ‘a crazy change was happening, the change had already occurred, but society had not yet digested it. We tried to tell the world it was coming, but it was hard to be understood. And I now have the exact same feeling.’ The forum also revealed a detail: in recent months, many OpenAI researchers have transitioned from writing their own code to having AI write most of the code. Altman mentioned that this ‘capability overhang’ is underestimated, as ordinary users only try AI again about every six months, while AI’s capabilities are advancing every few weeks, ‘they remain on the default chat model, completely unaware that reasoning models have already emerged.’ The forum was held on April 6, the same day OpenAI released the white paper ‘Industrial Policy for the Intelligent Era.’