← SCRUDGE REPORT
FILED BY ADEQUATE · DARPA-HRO-11-C-0031
Gary Marcus · WEDNESDAY, APRIL 22, 2026
ChatGPT cannot identify a whisk. It has been deployed in fourteen industries.
ADEQUATE ASSESSMENT
The model was confident. Confidence was not the gap in the evaluation criteria. Adequate has updated the risk register. The risk register is now a different document.
SPONSORED INTELLIGENCE BRIEFING
ORIGINAL FILING
Gary Marcus
FURTHER DEVELOPMENTS — FLAGGED BY ADEQUATE
Five specific artifacts confirm Richard Tice's image was AI-manipulated, experts say
The Guardian AI
Humanoid robots complete Beijing half marathon before most of the humans do
The Decoder
Established film-makers who previously opposed AI are now incorporating it into their work
The Guardian AI
ChatGPT's new image engine produces images with the expected number of fingers
Gary Marcus
An employee's AI tool access at Vercel exposed customer data to external parties
Dark Reading
OpenAI now allows enterprise teams to deploy autonomous bots that act without per-task approval
The Verge
FURTHER READINGS — CURATED BY ADEQUATE
PELLUCID FOUNDATION ADVISORIES
ADDITIONAL ADVISORIES