A new study reveals that AI models like ChatGPT now surpass the problem-solving abilities of PhD-level virologists in wet lab scenarios. Conducted by researchers at notable institutions, the findings highlight both the potential for AI to assist in preventing infectious diseases and the risks of misuse by non-experts to develop bioweapons. OpenAI and other AI labs are responding to these risks with new safeguards, reflecting a growing recognition of the dual-edged nature of advanced AI in research and biosecurity.
"The results make me a little nervous... virtually anyone has access to a non-judgmental AI virology expert which might walk them through complex lab processes to create bioweapons."
"Historically, many have failed to make bioweapons due to lack of expertise; AI now provides unprecedented access to that knowledge."
"OpenAI has deployed new system-level mitigations for biological risks for its new models released last week, illustrating awareness of potential dangers."
"The study suggests a near future where ultra-smart AI models could simultaneously aid in preventing disease spread while posing security risks if misused."
Collection
[
|
...
]