Anthropic Blocked Attempts to Use Claude for Bioweapon Research
A network news report puts on the record Anthropic's own findings that its AI was used to research a highly pathogenic bird flu variant — and that a former researcher warned of extinction-level risk within a decade.
1 / 5at 0:56 · Yank's paraphrase
A lab outside the U.S. gained access to Claude from an unsupported region, using the model to research a highly pathogenic bird flu variant and its potential transmission to humans.
A lab outside the U.S. gained access to Claude from an unsupported region, using the model to research a highly pathogenic bird flu variant and its potential transmission to humans.
In context
last year where the company says it disrupted potentially malicious use of its AI, including several examples of quote biological misuse. In one such instance, Anthropic says a lab outside the U .S. gained access to Claude from an unsupported region, using the model to help gained access to Claude from an unsupported region. research a highly pathogenic bird flu variant and its potential transmission to humans. Anthropic says in each instance it banned the associated user accounts, adding that they do not know if the users intended harm or
A lab outside the U .S. U.S. gained access to Claude from an unsupported region, using the model to help gained access to Claude from an unsupported region. research a highly pathogenic bird flu variant and its potential transmission to humans.
Receipt
Anthropic published a report saying it blocked several possible attempts to use its AI models to help develop biological weapons.
In context
outside this country was trying to use their A .I. to potentially develop a bioweapon. Here's Matt Rivers. Tonight, after one of its own former employees warned A .I. could kill all of humanity in a matter of years, Anthropic now publishing. an alarming new report saying it blocked several possible attempts to use its AI models to help develop biological weapons. The tech giant behind Claude detailing dozens of incidents since the end of last year where the company says it disrupted potentially malicious use of its AI, including
Anthropic now publishing. an alarming new published a report saying it blocked several possible attempts to use its AI models to help develop biological weapons.
Receipt
Anthropic says in each instance it banned the associated user accounts, adding they do not know if the users intended harm or were performing legitimate scientific work.
In context
the U .S. gained access to Claude from an unsupported region, using the model to help gained access to Claude from an unsupported region. research a highly pathogenic bird flu variant and its potential transmission to humans. Anthropic says in each instance it banned the associated user accounts, adding that they do not know if the users intended harm or were performing legitimate scientific work. The report comes after a former Anthropic and OpenAI researcher issued a dire warning about the possible dangers of AI. They could cause extreme
Anthropic says in each instance it banned the associated user accounts, adding that they do not know if the users intended harm or were performing legitimate scientific work.
Receipt
They could cause extreme havoc — hacking critical infrastructure, building extinction-level bioweapons.
In context
associated user accounts, adding that they do not know if the users intended harm or were performing legitimate scientific work. The report comes after a former Anthropic and OpenAI researcher issued a dire warning about the possible dangers of AI. They could cause extreme havoc, for example, hacking critical infrastructure, building extinction -level bioweapons. David, a growing number of lawmakers on Capitol Hill are calling for increased government oversight of AI development, including a proposed AI kill switch bill to
They could cause extreme havoc for example, — hacking critical infrastructure, building extinction -level extinction-level bioweapons.
Receipt
A proposed AI kill switch bill would halt models if they pose a catastrophic risk.
In context
were performing legitimate scientific work. The report comes after a former Anthropic and OpenAI researcher issued a dire warning about the possible dangers of AI. They could cause extreme havoc, for example, hacking critical infrastructure, building extinction -level bioweapons. David, a growing number of lawmakers on Capitol Hill are calling for increased government oversight of AI development, including a proposed AI kill switch bill to halt models if they pose a catastrophic risk. David, of course, we're going to stay on this issue so important and critical. Matt Rivers, thank you.
A proposed AI kill switch bill to would halt models if they pose a catastrophic risk.
Receipt
Keep going with this source
Create something
Turn this source into a post, a thread or a newsletter, every line checked against it.
See what you can create →Ask a question
Ask about anything in this video. We’ll try to provide an answer.