OpenAI is celebrating reaching its goal of building an "automated research intern" designed to help accelerate and support future model development.
The brand calls this “a system that can carry out well-defined research tasks under human direction, including tasks that would take a skilled researcher a few days.” It intends to ask researchers to use the tool to “further progress on deep learning and alignment, enabling iterative improvements.”
In a recent blog post, OpenAI noted how many of its own researchers now use coding agents to speed up their work and handle more complex tasks. Over the last few months, GPT-powered agents have been used for more internal tasks than ever before, freeing up OpenAI's researchers to run more experiments and develop future ideas.
CEO Sam Altman first announced the goal of developing an automated researcher last October, setting a deadline of September 2026. Altman also outlined the brand's next step, what he calls an “automated AI researcher", which OpenAI says it's on track to hit. That deadline is March 2028.
The blog post confirms that safety is a major concern for OpenAI's latest innovation, freely admitting it doesn’t yet know how to run its ultimate automated AI researcher safely. That’s especially pertinent given ongoing security incidents with other OpenAI tools.
This news comes at a time when OpenAI is navigating a series of what it calls “misalignments," including the Hugging Face hack and the recently confirmed incident where autonomous agents edited entries over 15,000 times on the German, Wikipedia-like service DseWiki.
OpenAI says, “Whenever we find that proceeding would pose an unacceptable safety risk, we will respond appropriately including by slowing or stopping our development or deployment of systems we find ourselves unable to sufficiently safeguard."
Last week, OpenAI introduced its new Astra model, which is now available to subscribers worldwide. The brand called it the "world's most intelligent and aligned model."
Disclosure: Ziff Davis, PCMag's parent company, filed a lawsuit against OpenAI in April 2025, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.