Artificial intelligence has already transformed how people write, search for information and develop software.
The next major transformation may happen somewhere less visible.
The laboratory.
AI company Anthropic has introduced a new system designed to allow AI agents to interact with scientific and engineering equipment, including microscopes, robotic arms, lasers and laboratory instruments.
The Model Hardware Standard is designed to give AI systems a way to communicate with physical devices and potentially carry out parts of scientific experiments.
The idea could represent a major change in how scientific research is performed.
Traditionally, scientific discovery involves a combination of human creativity and physical experimentation.
Researchers develop a hypothesis.
They design an experiment.
They operate equipment.
They analyse the results.
They decide what should happen next.
Increasingly, AI could become involved in several stages of that process.
An AI system might analyse previous research and identify possible experiments.
A robotic system could carry out those experiments.
The AI could then examine the results and suggest what should happen next.
In theory, this could create laboratories capable of operating continuously and exploring scientific questions at speeds difficult for human researchers to match.
Anthropic has begun testing the technology with research organisations and companies, including institutions involved in biological research, engineering and quantum technology.
The potential applications are enormous.
Drug discovery is one obvious example.
Developing new medicines often requires testing huge numbers of chemical compounds.
AI could help researchers identify promising candidates while automated laboratories test them.
Materials science could also benefit.
Scientists could potentially use AI to search for new materials with properties useful for batteries, clean energy or electronics.
The technology could even influence manufacturing.
AI agents capable of interacting with physical equipment could help factories monitor systems and adjust processes more intelligently.
But the development also raises important questions.
Science is not simply a process of collecting data.
Researchers must interpret results, question assumptions and recognise when something unexpected is happening.
An AI system may be capable of running experiments.
That does not necessarily mean it understands the significance of what it discovers.
Human scientists will remain essential.
The more realistic future may involve collaboration.
AI could handle repetitive tasks and analyse enormous datasets.
Humans could focus more heavily on creativity, judgement and the interpretation of discoveries.
However, giving AI systems control over physical equipment introduces new risks.
A mistake in an ordinary chatbot conversation may be inconvenient.
A mistake involving a chemical laboratory or industrial machine could have physical consequences.
That is why companies developing these systems are placing increasing emphasis on permissions, safety controls and human oversight.
The future laboratory may therefore look very different from the laboratories of today.
Scientists may work alongside AI systems capable of processing information faster than any human.
Robotic equipment may carry out experiments around the clock.
Research may become increasingly automated.
And discoveries that once required years of trial and error could potentially happen much faster.
The World Economic Forum’s 2026 emerging technologies report also reflects this wider trend, highlighting scientific advances across energy, materials, health and computing that are beginning to move from research towards real-world deployment.
The most important change may not simply be that AI becomes another scientific tool.
It may become part of the scientific process itself.
Humanity has always developed new tools to understand the world.
Telescopes expanded what humans could see.
Microscopes revealed worlds too small for the naked eye.
Computers allowed scientists to process enormous amounts of information.
Artificial intelligence could become the next great instrument of discovery.
Not because it replaces scientists.
But because it may help them discover things they would never have had time to find alone.
