The AI biosecurity letter everyone's covering wrong
OpenAI and Anthropic want Congress to regulate DNA synthesis companies, not themselves. That's the real story here.
Crédit photo: Image via Source article. Used under fair use for news commentary. · source
Most coverage of this week's open letter from AI executives frames it as "AI leaders want regulation." And sure, technically that's true. But I think we're missing something important here.
The letter, signed by OpenAI's Sam Altman, Anthropic's Dario Amodei, Microsoft's Mustafa Suleyman, and a bunch of scientists, isn't asking Congress to regulate AI companies. It's asking Congress to regulate DNA synthesis companies. That's a pretty significant distinction that's getting lost in the headlines.
What they're actually asking for
The core request is straightforward: require companies that sell synthetic DNA and RNA online to screen orders for dangerous sequences. Right now, you can apparently order genetic material that could theoretically be used to engineer pathogens, and there's no mandatory screening process to catch that.
I should be honest here, I'm not a biosecurity expert. But the basic argument makes sense. AI tools are getting better at helping people understand how to design biological agents. If those same people can then order the raw genetic materials without anyone checking what they're building, that's a gap.
The letter calls it a "biosecurity gap that could help trigger a global pandemic." Strong language, but coming from people who presumably understand what their own models can do.
Why this is interesting (and a little convenient)
Here's where I get a bit skeptical, honestly. This letter lets AI companies look responsible while pushing the regulatory burden onto a completely different industry. DNA synthesis companies would need to implement screening. AI labs would just... keep doing what they're doing.
À lire aussi
More in Policy
At the Northeast National Security Conference, officials called for rapid innovation pathways while manufacturers grapple with the harder problem of actually building things domestically.
James Chen · 16 hours ago · 4 min
Wall Street's celebrating AI-fueled gains while missiles fly over the Middle East. I've seen this disconnect before, and it never ends well.
Mark Kowalski · 2 days ago · 5 min
Google's promising to replenish more water than it uses by 2030. I've heard promises like this before.
Mark Kowalski · 2 days ago · 6 min
The administration killed Biden's AI regulations, then quietly built something that looks... kind of similar? I tried to make sense of it.
