AI Making Viruses Still Needs Real-World Labs, Researchers Say

Debate about AI possibly leading to human extinction has heated up recently, but several researchers say there is still a big gap between AI's digital abilities and real-world harm.

Google senior research engineer Anselm Levskaya used AI-designed viruses as an example. Even if AI generates a virus sequence, it still must be assembled in a lab and turned into virus particles, then go through lots of screening and testing. These depend on complex biology, not something that can be designed just by sitting at a computer. UAE Mohamed bin Zayed University of AI president Xing Bo said generating a virus blueprint and actually making a virus are completely different things.

Researchers stress that for AI to turn a plan into real harm, it must reach physical equipment, money, or people. Cohere CEO Aidan Gomez said the debate over AI existential risk has slid too far into science fiction. Georgia Law assistant professor Mateusz Błażyk warned that even if extreme scenarios are doubted, people should not ignore more realistic risks like AI worsening cyberattacks and surveillance.