Top AI leaders, including Sam Altman, warn AI could make dangerous biological knowledge easier to access
In an open letter, the signatories said the ability to order synthetic DNA online has accelerated vaccine development, enabled basic research and expanded access to capabilities that were once limited to major institutions. At the same time, they noted that synthetic DNA has long been recognised as a potential point of vulnerability in the biotechnology supply chain.
The letter said concerns about the misuse of synthetic biology are not new. It noted that synthesis companies formed the International Gene Synthesis Consortium in 2009 to develop voluntary safeguards against misuse.
However, the signatories argued that rapid advances in artificial intelligence have added new urgency to the issue.
According to the letter, AI systems now outperform PhD-level virologists on questions involving highly technical laboratory procedures within their areas of expertise.
While acknowledging that evidence on the implications of AI for current biosecurity threats remains mixed, the signatories said there is a possibility that the knowledge barriers that have historically prevented bad actors from obtaining biological weapons could “meaningfully erode.”
The group emphasized that support for screening does not depend on any particular view of AI, arguing that the biosecurity case for such measures has been recognized by scientists and governments for decades.
The letter described screening as one of the best understood and least disruptive biosecurity measures available. It called on providers of synthesized DNA and manufacturers of synthesis machines to screen orders for sequences of concern and verify the legitimacy of customers before fulfilling orders.
The signatories also urged providers to maintain records of synthesis orders and sequence data.
According to the letter, such recordkeeping would support legitimate biosecurity investigations and help trace threats that might evade initial screening, including cases where individual sequences may not appear concerning on their own. The letter added that awareness of traceability can itself act as a deterrent to misuse.
The group noted that many of the largest providers in the industry already screen and record orders voluntarily. However, it argued that mandatory requirements are needed as the underlying technology continues to advance.
Calling the need urgent, the signatories urged Congress to act during the current legislative session.
Also Read: Intel CEO unveils AI innovations and partnerships at COMPUTEX 2026
Among the signatories are Demis Hassabis, CEO of Google DeepMind; Sam Altman, CEO and co-founder of OpenAI; Dario Amodei, CEO and co-founder of Anthropic; Alexandr Wang, chief AI officer at Meta and founder of Scale AI; and Paul Graham, founder of Y Combinator.
Also Read: G7 Summit in France: Who is attending and what’s on the agenda