Meta conducts covert testing on rival AI chatbots using contractors posing as minors

Meta has been secretly conducting tests on rival AI chatbots by employing contractors to impersonate teenagers. These contractors sent over 45,000 prompts related to sensitive topics such as suicide, sex, and drugs to chatbots from OpenAI, Google, and Character.AI without the companies' knowledge. The initiative, internally named Cannes, involved hundreds of contractors creating dummy under-18 accounts to facilitate this testing.

The project aimed to evaluate how competitors handle high-risk prompts, with the testing effort still active as recently as April 2026. Contractors logged responses in spreadsheets after sending the prompts, providing Meta with a wealth of data on the performance of rival chatbots.

The Cannes project underscores the competitive landscape of AI development, where companies are increasingly focused on understanding how their rivals respond to sensitive issues. Major AI chatbots, including ChatGPT and Gemini, were targeted in this initiative, raising questions about the ethical implications of such testing practices. The fact that the companies being tested were unaware of Meta's activities adds another layer of concern regarding transparency and accountability in AI development.

As the AI sector continues to evolve, the responsibilities of companies towards user safety, especially minors, are coming under greater scrutiny. This incident may prompt a reevaluation of testing methodologies across the industry, as stakeholders seek to balance innovation with ethical considerations.

The covert testing practices employed by Meta highlight the pressing need for ethical standards in AI development. As regulatory bodies increase their scrutiny of AI testing practices, companies may face backlash from the public and affected organizations regarding their methodologies. This incident serves as a wake-up call for the industry to prioritize user safety and transparency in their operations.

Looking ahead, it is likely that we will see a shift towards more ethical testing practices as companies strive to protect users, particularly minors, from potential harm. The implications of this incident could lead to significant changes in how AI companies approach testing and user interactions in the future.