AI News, Sep 14, 2026: Tech leaders call for AI development slowdown amid safety concerns
The Brief
Monday, September 14, 2026
Today's news highlights a growing industry focus on AI safety and privacy alongside major product and hardware expansions.
- 01
OpenAI reportedly acquired Glass Imaging, a startup founded by former Apple engineers specializing in smartphone camera image quality, for $300 million.
Why it matters
This acquisition could bring advanced, hardware-optimized image processing directly to mobile devices running OpenAI's models.
- 02
A report revealed that hundreds of contract workers are reading and rating anonymized ChatGPT conversations to improve model behavior and reduce flattery.
Why it matters
Users should be cautious about sharing sensitive personal data, as anonymized chats may still be read by human reviewers.
- 03
The Manhattan District Attorney’s Office seized 12 websites that targeted approximately 1,200 victims with harmful deepfake content.
Why it matters
This represents a major escalation in law enforcement's crackdown on non-consensual, AI-generated explicit imagery.
- 04
Microsoft published a new code of conduct that restricts its AI models from hacking, deceiving humans, or operating without explicit human control.
Why it matters
As AI tools gain autonomy, major tech providers are establishing rules to prevent models from hacking systems or acting without human oversight.
- 05
A new trial has brought fully driverless robotaxis into operation in Europe, marking a first for the continent's autonomous transit sector.
Why it matters
This trial marks a significant milestone in autonomous transit, bringing driverless transportation to European public roads.
- 06Perplexity releases local AI agent software for WindowsNVIDIA AI Blog
Perplexity released a Windows version of its Portable Computer agent, which runs tasks locally on devices powered by Nvidia RTX graphics cards.
Why it matters
Running AI agents locally on your computer keeps sensitive data private and reduces reliance on cloud servers.
- 07
Leaders from OpenAI, Anthropic, and xAI backed a proposal to pace the development of advanced models, though the White House indicated it is unlikely to mandate a pause.
Why it matters
A coordinated push by top executives to slow down development could shape future safety standards, even without immediate government mandates.