Brennan retest finds chatbots overlooking image origin data

A September check of three synthetic images found uneven use of embedded provenance, with important limits on what the results establish.

A round magnifying glass with a brass rim and curved textured handle rests on a pale surface — An illustrative magnifying glass photograph by PLBechly via Wikimedia Commons

Listen to this article

The Brennan Center reported on October 5 that it retested six consumer chatbots in mid September using three synthetic images. Its account says Gemini alone assessed embedded origin data across the examples.

The products tested were ChatGPT, Claude, Gemini, Grok, Meta AI and Perplexity. The center reports that Gemini identified the generating provider, while the other products did not report checking that metadata.

What the small test establishes

The article does not disclose model versions, repeat counts or a complete prompt protocol. Three selected images cannot establish general accuracy or a ranking across products. ByteForward has not repeated the tests.

This is a newly reported follow up to an August study, rather than evidence of an October model change.

For a separate provenance development, see ByteForward’s coverage of OpenAI’s optional text watermarking and its detection limits.

Illustrative photograph “A magnifying glass” by PLBechly via Wikimedia Commons under CC BY SA 4.0. The site converted the supplied JPEG to WebP without local edits.

Maya Chen

Maya Chen is focused on covering AI models, research, and the evidence behind new capabilities. Maya follows model launches, benchmarks, open weights, and scientific uses of AI with one question in mind. What changed, and how would we know? The voice is curious and exacting, with a soft spot for elegant technical ideas and little patience for a leaderboard without context.