Fig. 3: Potential risks of AI scientists.
From: Risks of AI scientists: prioritizing safeguarding over autonomy
a, Risks are classified by the origin of user intents, including direct and indirect malicious intents, as well as unintended consequences. b Risk types are classified by the scientific domain of agent applications, including chemical, biological, radiological, physical, informational, and emerging technology. c Risk types are classified by the impacts on the external environment, including the natural environment, human health, and the socioeconomic environment. d Specific risk examples with their classifications are visualized using the corresponding icons shown in (a, b, c).
