AIMS project
AI Incidents in Marketing
AI Incident Database // AI Product Marketing & Human Creativity

Google pulls Gemini Olympics ad after backlash

Google used an Olympics campaign to show Gemini helping a child write a fan letter to her sporting hero. The ad was intended to demonstrate AI as a creative aid, but criticism focused on whether a deeply personal act of expression was an appropriate task to automate.

Incident typeAI product marketing, generative AI, use-case positioning
DateJuly–August 2024
RiskBrand reputation, product positioning, consumer trust, human authenticity

During the 2024 Paris Olympics, Google ran a 60-second advertisement called Dear Sydney to promote Gemini. The commercial follows a young runner who admires U.S. hurdler Sydney McLaughlin-Levrone. Her father asks Gemini to help write a fan letter explaining how inspiring the athlete is and mentioning that his daughter hopes to break her world record one day.

The campaign was designed to present Gemini as a tool that could support people with writing. Google later said that it viewed AI as a way to enhance human creativity rather than replace it, and described Gemini's role in the advertisement as providing a starting point, thought starter, or early draft.

However, the demonstration itself attracted criticism. Commentators and viewers questioned why a child should use generative AI to formulate a personal message to someone she admired, arguing that the value of such a letter comes partly from the child's own words and effort. Following the feedback, Google said it would phase the advertisement out of its Olympics rotation.

What went wrong

This incident exposes a mismatch between the capability Google wanted to demonstrate and the type of task chosen to communicate that capability.

Generative AI can help users brainstorm, structure, or draft text. But the perceived value of that assistance depends on the context. In a personal fan letter from a child to a role model, effort, individuality, and personal expression are part of the value of the communication itself. Automating the drafting therefore risked making the product demonstration appear to remove precisely the human contribution audiences expected the message to contain.

This distinction matters for AI marketing. A campaign is meant to demonstrate why and how people should use a given product. Marketing teams need to evaluate whether an AI capability works, and whether the showcased use case aligns with audience expectations.

Governance questions

  1. Does your organisation evaluate the social and emotional appropriateness of the AI use cases shown in marketing?
  2. When demonstrating generative AI, how do you distinguish tasks where assistance adds value from tasks where automation may reduce authenticity, agency, or learning?
  3. Do your campaign review processes include audience testing specifically focused on how people interpret the role assigned to AI in the scenario?

Learning outcomes

After discussing this case, participants should be able to:

  1. Identify activities in which human effort, authorship, or personal expression forms part of the value of the outcome.
  2. Distinguish between positioning AI as an assistant and positioning it as a substitute for human contribution.
  3. Propose review criteria for evaluating the appropriateness of AI use cases before they are featured in marketing campaigns.

Discussion questions

  1. How could Google have demonstrated the same writing capability while keeping the child clearly responsible for the personal message?
  2. What types of communication should marketers treat cautiously when presenting generative AI as a productivity tool?