00:00It's Benzinga bringing Wall Street to Main Street.
00:02OpenAI said it cannot rule out that its upcoming AI model Astra has critical cyber security
00:08capabilities after preliminary evaluations showed it could perform increasingly sophisticated
00:14cyber tasks autonomously, according to Reuters. Under the company's safety guidelines,
00:19that threshold includes autonomously exploiting severe zero-day vulnerabilities
00:24or carrying out complex attacks against highly secure targets.
00:28OpenAI said preliminary evaluations and outside expert assessments showed Astra may have
00:34advanced enough autonomous cyber capabilities that the company cannot rule out a critical
00:39capability level. OpenAI paused internal Astra activities that do not meet strengthened security
00:45requirements and will move development into isolated testing environments with restricted
00:50network access and sandboxed execution. The company plans to work with government agencies and select
00:56AI safety organizations to test Astra. OpenAI said Astra was not involved in the July hack
01:02targeting hugging face. For all things money, visit Benzinga.com.
Comments