Dr. Richard Moulange
speaker
1,784 appearances
1 recordings
1 series
first heard Mar 2026
last heard 31 Mar
Dr. Richard Moulange’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 1 in all, peaking in Mar 2026 with 1.
Appearances
have a tool that will take a sequence and then convert it into a, highlight a good function, say, and say, yes, this thing helps increase immune resistance to something.
Then you also have a tool that can take a sequence and say, oh, this part of the sequence is a virulence factor.
This is actually a really, really important thing for causing more harm.
And even more than that, it'll be the other way around, which will be function to sequence, the inverse problem, where you say, I want to cure disease.
What sequence should I use?
I want to cause disease, what sequence should I use?
So we as a society should be really, really careful about bringing that capability into the world.
But it is coming because the commercial relevance is staggering.
And so instead, we as the governance and the risk-aware community should be thinking, given that capability is going to come into the world,
How can we steer it?
What are the defensive beneficial applications that we would wish to have?
How can we make sure we incorporate those early?
Because misuse applications will also arise.
Yes, so there are three sort of main types of interventions that I hear discussed about dealing with the AI bioproblem.
One is managed access.
Can we have it so that not everyone is able to access advanced AI assistance, whether troubleshooting from LLMs or biological design from specialized tools, so that then malicious actors can't get them?
Nothing happens.
And we don't have a pandemic, an AI-enabled pandemic.
The second one is guardrails, safeguards.
Even if somebody scary can access a tool, maybe the tool itself will refuse to do this.
Showing 641–660 of 1,784 · page 33 of 90
← Previous
Next →