OpenAI released ChatGPT 5.6 after complying with Trump administration requests to restrict initial access and conduct additional testing. Similar restrictions were placed on Anthropic models, resulting in a temporary export ban. CNA reported the separate launch of ChatGPT Work on July 9 while disputing its connection to the new model.
“OpenAI releases latest ChatGPT model after delay over White House cybersecurity concerns”
Read at The Guardian →No right-leaning sources covered this
The sequence shows the value of public oversight on frontier AI but also the limits of ad-hoc White House interventions.
“Necessity of external checks on corporate power versus risks of opaque political processes”
The administration’s targeted restrictions addressed national security risks before wider release while allowing eventual public access.
“Protection of sensitive technologies from foreign exploitation”
Government intervention delayed private product releases and substituted bureaucratic judgment for market adoption.
“Regulatory capture and expansion of administrative power over private R&D”
All three views accept unsubstantiated cybersecurity premises and overlook commercial timing and sourcing limitations.
“Shared assumptions about the nature of the review process and absence of scrutiny on evidence or competitive effects”
Ratings by MBFC