One of Google DeepMind’s top researchers says AI agents are beginning to learn and adapt with far less direct human supervision.
In a new post on X, Google DeepMind’s director of research, Edward Grefenstette, says his team is seeing progress toward AI systems that can improve themselves while operating in real-world environments.
At a high level, Grefenstette says AI agents are starting to learn without the need for constant human guidance or carefully labeled training data. Instead of being trained only in controlled settings, these agents can learn while interacting with the world, adjusting their behavior as they go.
“Broadly, we’ve been making good progress with regard to how open-ended agents can learn ‘in the wild,’ with less human intervention in their learning process, while still ensuring they remain aligned with human behaviors and interests.”
In practical terms, this means AI systems that can explore new situations, figure out what works and what does not and still stay within boundaries that humans consider safe or useful.
Grefenstette also points to improvements in how efficiently these agents learn. Traditionally, AI systems have required enormous amounts of data to improve. He says newer approaches allow agents to adapt much faster, more like humans do, meaning an AI agent could potentially learn a new task after seeing only a small number of examples, rather than millions.
“We’ve also made some progress in terms of the actual learning process itself, allowing open-ended agents, at the instance level, to learn and adapt with human-like data efficiency. This potentially points at a broader way of improving agents at scale, which we are working on.”
If successful, the approach could accelerate the development of autonomous AI systems capable of operating and collaborating continuously in real-world settings.
Last month, Google said that 2026 will be the year when AI agents transition from assisting humans to completing entire workflows while collaborating with each other.
Disclaimer: Opinions expressed at CapitalAI Daily are not investment advice. Investors should do their own due diligence before making any decisions involving securities, cryptocurrencies, or digital assets. Your transfers and trades are at your own risk, and any losses you may incur are your responsibility. CapitalAI Daily does not recommend the buying or selling of any assets, nor is CapitalAI Daily an investment advisor. See our Editorial Standards and Terms of Use.