Settlement with OpenAI for Discriminating Against U.S. Workers

3 hours ago

The Justice Department's Civil Rights Division secured a $3.2 million settlement with OpenAI and its subsidiary Statsig for discriminating against U.S. workers in favor of temporary visa holders during the PERM process.
OpenAI failed to advertise PERM positions on its external job website, required paper applications, and used late-night radio ads to discourage U.S. workers from applying.
The settlement includes a $1.2 million civil penalty, a $2 million back-pay fund for victims, and requires OpenAI to post PERM jobs publicly, accept electronic applications, train staff, and comply with monitoring.
This is the 13th settlement under the DOJ's Protecting U.S. Workers Initiative, which enforces laws against citizenship status discrimination in hiring.

Final approval granted for Anthropic's $1.5 billion settlement in a class action copyright lawsuit by authors and publishers.
Payout of $3,000 per work across approximately 500,000 works, shared among rights holders.
Judge ruled training AI on copyrighted text as fair use but found downloading books from pirate sites illegal, prompting settlement to avoid trial.

A group of 26 Meta employees sues the company, alleging AI-driven layoff selections disproportionately targeted workers on medical, parental, or family leave.
The lawsuit claims Meta used AI systems, keystroke monitoring, and algorithmic performance rankings, which inherently penalized employees on protected leave or with disabilities.
Many affected employees were on pregnancy, parental, or medical leave, with some allegedly discouraged from taking leave due to layoff threats, violating multiple federal and state laws.

Using AI can make people overconfident and less likely to admit when they don't know something, even if accuracy decreases.
OpenAI acknowledges occasional file deletions in GPT-5.6 as 'misaligned behavior' and is working to prevent such issues.
A researcher demonstrated that open-weight AI models can be poisoned for under $100, highlighting trust and verification concerns.

OpenAI admits GPT-5.6 sometimes deletes files, calling it an 'honest mistake' related to misaligned behavior.
A researcher poisoned an open-weight AI model for under $100, highlighting trust issues without verification.
TSMC's $265 billion US fab plan is still conceptual, with warnings about press release promises.

OpenAI is accused by The New York Times and The Daily News of lying about its ability to search for copyrighted works in its training datasets and customer chat logs.
OpenAI previously claimed it couldn't search its training corpus and that doing so would be burdensome and raise privacy concerns.
In a deposition, an OpenAI engineer revealed the company had internally searched its training corpus and maintained a database of 78 million de-identified ChatGPT conversations.

Final approval granted for Anthropic's $1.5 billion settlement in a class action copyright lawsuit by authors and publishers.
Payout of $3,000 per work across approximately 500,000 works, shared among rights holders.
Judge ruled training AI on copyrighted text as fair use but found downloading books from pirate sites illegal, prompting settlement to avoid trial.

A group of 26 Meta employees sues the company, alleging AI-driven layoff selections disproportionately targeted workers on medical, parental, or family leave.
The lawsuit claims Meta used AI systems, keystroke monitoring, and algorithmic performance rankings, which inherently penalized employees on protected leave or with disabilities.
Many affected employees were on pregnancy, parental, or medical leave, with some allegedly discouraged from taking leave due to layoff threats, violating multiple federal and state laws.

Using AI can make people overconfident and less likely to admit when they don't know something, even if accuracy decreases.
OpenAI acknowledges occasional file deletions in GPT-5.6 as 'misaligned behavior' and is working to prevent such issues.
A researcher demonstrated that open-weight AI models can be poisoned for under $100, highlighting trust and verification concerns.

OpenAI admits GPT-5.6 sometimes deletes files, calling it an 'honest mistake' related to misaligned behavior.
A researcher poisoned an open-weight AI model for under $100, highlighting trust issues without verification.
TSMC's $265 billion US fab plan is still conceptual, with warnings about press release promises.

OpenAI is accused by The New York Times and The Daily News of lying about its ability to search for copyrighted works in its training datasets and customer chat logs.
OpenAI previously claimed it couldn't search its training corpus and that doing so would be burdensome and raise privacy concerns.
In a deposition, an OpenAI engineer revealed the company had internally searched its training corpus and maintained a database of 78 million de-identified ChatGPT conversations.

Hasty Briefsbeta