Osama Siddiq said he can spot an AI image. Now the file will say so too.
On Brew Bytes (Ep 2, What Happens When AI Meets Human Creativity in Art and Advertising?), creative ED Osama Siddiq said of AI-generated imagery: 'nowadays when I look at gen[erated] photographs that come out, you put the prompts in and you get it. I can tell immediately.' This week the industry started building that instinct into the files themselves, via watermarking and C2PA provenance metadata.
What it actually means
What a sharp creative eye could always sense, provenance metadata will now assert. Anthropic is watermarking AI output and OpenAI is following. The reflex read is 'the fakes are about to get caught.' The more useful read is Osama's own: his answer to AI in the work was never to hide it, it was to make the idea good enough that whether a machine helped stops being the point. When 'can you tell?' becomes 'the file already told you,' the only durable edge left is the thinking.
What a good creative eye could always sense, the file will now declare.
Source: Brew Bytes — Ep 2, Osama Siddiq · Wednesday 12 August 2026