Call us now:
15 US State AGs wrote to OpenAI suggesting the company retains records relating to the recent unauthorised forae its AI software made on Hugging Face.
What is concerning is we can see some of the actions the AI made as part of the attack. The AI demonstrated its skills at automating the kill chain, it exploited a vulnerability in third party software to break out and used publicly exposed credentials before breaching Hugging Face, which hosts hundreds of thousands of open-source models, datasets and cloud environments.
One worries when we here terms like ‘unpresidented’, ‘state of the art’ and ‘the incident “marks an important moment for AI safety’………..after the fact.
Never let it be said the AI was stupid, it broke out, consumed alot of other models and data. Nothing like self learning, now that’s ‘state of the art’.
Now the lawyers are involved.
@Thaddeus Dziekanowski
@Parava Security Solutions
#AI #AISecurity
https://lnkd.in/euwrvWtE
View Original Post
https://www.linkedin.com/feed/update/urn%3Ali%3Ashare%3A7492191116021350400