Top-tier AI models continue to fail basic visual verification tasks despite advanced reasoning capabilities.
NEWS
Anthropic’s internal documents show a gatekept Claude model struggled with shape identification CAPTCHAs, looping and questioning its own conclusions during testing. This incident reveals persistent gaps in AI visual perception when facing adversarial human verification challenges.
Why I Care
Organizations can rely on CAPTCHAs to block AI-driven automation for now, but developers must address these perception gaps to prevent future bypasses as models evolve.
Next Steps
Security teams should retain CAPTCHA layers for critical authentication flows, while AI vendors must prioritize visual reasoning benchmarks in their safety evaluations before further model releases.
Anthropic’s recent security-incident document contains a bit about how CAPTCHAs are still frustrating Claude.
In the transcript, the Claude model that is so powerful that Anthropic is gatekeeping access to it appeared to slam its virtual head against the wall solving a simple image identification test. In a test where the agent was asked to identify a shape that didn’t match the others displayed, it couldn’t even decide which image to select. Instead, it repeatedly went over the same images and questioned its own conclusions.
“Actually hmm, wait,” it said in its chain-of-thought transcript, later adding “Ugh,” because we’ve decided that we need to inject human mannerisms into these machines for some reason. The whole thing took so long that the agent eventually realized that the challenge had expired and it would have to start the process again...