AI Resilience Report forPhysical Scientists, All Other

Physical scientists study non-living materials and energy (for example physics, chemistry, earth and space). Some parts of their work are already aided by machines and AI. For example, research labs now use robots to run long experiments without tiring, so scientists can spend more time planning tests and checking results ^[1].

Astronomers note that only AI can help sort through the huge amounts of telescope data they collect ^[2]. Similarly, earth and environmental scientists use AI to analyze satellite images (for example to track mining changes or deforestation) ^[3]. These tools handle routine data tasks, but people still decide what experiments to do and how to interpret results ^{[1] [2]}.

In short, AI helps with big data and repetitive work, but human researchers remain in charge of creative and complex thinking.

Many scientists say AI tools are already saving them time and money, which encourages labs to try them ^[4]. But how quickly AI spreads in this field depends on practical issues. AI systems and lab robots can be expensive and need to prove they work well.

Industry experts note that new robots must become cheap enough, fast enough, and reliable enough before they bring big value ^[5]. Some researchers also warn against depending too much on AI ^[4]. On the positive side, science jobs are expected to grow faster than average ^[6], so there should be demand for human scientists.

In the end, a scientist’s creativity, judgment and problem-solving skills – like designing an experiment or understanding nature – are hard to automate and will stay very valuable ^{[1] [4]}.