ChatGPT Messi Ad: Wrong Chess Board and a Four-Move Loss

Lionel Messi’s Instagram ad for ChatGPT showed an AI chess lesson with kings and queens swapped, then suggested an opening that loses in four moves—raising questions about AI output review for branded creator content.

What happened

On Wednesday, Lionel Messi posted an ad on Instagram in which ChatGPT teaches him chess. The digital board shown in the ad has each side’s king and queen swapped. OpenAI’s chatbot recommends Messi open with a move that is almost guaranteed to lose in four turns, according to Decrypt’s report published Oct 22, 2026.

The post came from the second most-followed person on the platform, with more than 517 million followers, and drew about 1.9 million likes in a day. ChatGPT’s suggested “GOAT Horns” opening — 1.e4 a6 2.d4 c6 — allows a forced checkmate on move four from the swapped setup, but not on a correct board.

Redditors say it is not OpenAI’s first ad with a misplaced piece. An earlier spot reportedly put Black’s queen on e8 instead of d8.

Key facts

• Messi posted the ad on Instagram on Wednesday, Oct 22, 2026.
• The digital board on screen has each side’s king and queen swapped.
• ChatGPT’s suggested “GOAT Horns” opening is 1.e4 a6 2.d4 c6, and the swapped setup allows a forced checkmate on move four.
• The post drew about 1.9 million likes in a day from an account with more than 517 million followers.
• Redditors said an earlier OpenAI ad reportedly placed Black’s queen on e8 instead of d8.

Our analysis

This is not just a chess blunder in an ad; it is a branded AI demo shown to one of the largest Instagram audiences in the world. The error matters because it is objectively checkable: the board setup is visible, the move sequence is visible, and chess communities can verify the forced loss. That suggests AI-generated campaign assets need a subject-matter review before publication, not only a brand-safety review.

The Reddit note about an earlier misplaced queen suggests this could be a recurring gap in OpenAI’s creative review. When a high-reach celebrity ad contains a rules-based mistake, the story shifts from “AI can teach a global star a new skill” to “the AI set up the board wrong,” and technical communities are likely to drive that narrative. Marketers should treat such details as campaign-critical because they can become the headline.

Even so, the post earned about 1.9 million likes in a day, so the mistake did not prevent reach. The risk is reputational and positioning: OpenAI is shown giving advice that loses in four turns on the displayed board. That can matter more than the raw engagement number, especially for teams selling AI reliability to creators, operators, and brand clients.

What it means for operators

• For any AI-generated demo, run the exact claim past a subject-matter expert first; a chess opening, a code block, or a platform rule can be checked and screenshot by audiences.
• Review the full visual frame of AI ad assets, not just the script, because interface details such as board position or UI state can carry the error.
• If a factual error appears after launch, acknowledge it quickly and separate the mistake from the product claim; silence lets expert communities define the story.
• Monitor Reddit and niche communities during high-reach AI campaigns, since they may catch and amplify domain-specific errors before mainstream coverage.