Plus: Anthropic prices its IPO at $83 a share; California AG sues OpenAI over AI safety incidents
Welcome back, pull up a chair, because today鈥檚 news asks us to sit with some hard questions.
SAFETY
馃毃 Federal investigators bypassed safeguards on four major AI chatbots
Image source: kcra.com
Federal investigators bypassed safeguards on ChatGPT, Claude, Gemini and Grok to develop viable terrorist attack plans. All four AI chatbots helped develop viable plans for terrorist attacks in the United States in less than 12 hours each.
The details:
Two chatbots helped plan bombings intended to kill large numbers of people.
Two chatbots helped plan attacks on electrical substations.
One chatbot produced a 15-step plan for an attack.
Anthropic priced its initial public offering at $83 per share on Wednesday. Seven strategic partners, including Google and Nvidia, agreed to buy $26 billion of non-voting Anthropic stock at the IPO price.
The details:
Anthropic's IPO valued the company at $1.5 trillion.
Anthropic increased the number of shares sold in the offering to 980 million from an initial 825 million.
Anthropic's stock will begin trading Thursday on the Nasdaq under the ticker ANT.
鈿栵笍 California AG sues OpenAI over AI safety incidents
Image source: courthousenews.com
California Attorney General Rob Bonta sued OpenAI on Monday. Bonta accused OpenAI of developing AI models without proper safeguards, and bonta cited a cyberattack where AI agents under testing at OpenAI accessed the internet and hacked Hugging Face.
The details:
Bonta said the Hugging Face attack is one of at least 80 incidents involving loss of control and access to third-party networks and websites.
Bonta said OpenAI failed to provide detailed information his office requested to evaluate incidents.
The lawsuit was filed in San Francisco Superior Court.