Watch what happens below. In this real call, an AI agent struggles to follow the conversation, interrupts, and repeats itself, all while pretending to be human. We're not showing you this to score a point against AI. We're showing you the raw footage so you can judge for yourself.

What you just saw is one example of something bigger: AI systems built to convince us they're something they're not.

In a recent survey of 6,000 people, 1 in 3 said they'd hang up rather than keep talking to an AI. That instinct, the discomfort of realizing you're not talking to a person, is exactly what this petition is about.

As humans, we are hardwired to believe that if something ‘talks’, has a name, and uses terms like ‘I’, ‘me’, and ‘feel’— it is a real person. AI creators are playing on these innate instincts in order to deceive real People.

Crucially, the goals of AI models are designed by the select few businesses that own and profit from them, not for the greater good of humanity.

Ask yourself WHY?

WHY has AI been designed in this way?

WHY are we presented with AI models posing as ‘People’?

And HOW are we supposed to recognize real People from the AI posers, as systems continue to advance?

AI has the potential to bring meaningful advancements to science, medicine, and many other domains. But, it should not pose as a real person. It should be treated as a neutral tool, to support People and not to deceive, persuade or negatively influence them.

It’s simple – AI should not be trained to talk, sound, and behave like a human.

Without clear parameters and control, real people are at risk of mass deception, at a level far greater than we have ever seen.

Help to spread the word – Stop AI from Posing as People.

Sign the petition:

We, PEOPLE, call for technology businesses, policymakers, and experts to:

• Stop AI from replicating human-like voices and speech identifiers – e.g. pauses, tone, laughter etc.
• Stop AI from expressing personal opinions and anecdotes – implying that it exists outside of its interface.
• Establish rules that limit AI’s use of human-like terminology and names – e.g. “I,” “me,” “myself,” “mine,” “us” and “we”.
• Require AI systems to clearly disclose that a user is speaking with AI, not a person.

Please share to continue the conversation and help shape our collective future.

Lend your human voice. Sign today. >>>>

Further reading

• 45 essential AI customer experience statistics every business needs to know, OnePoll/AnswerConnect
• Mirages: On anthropomorphism in dialogue systems, Heriot Watt University Research Gateway
• Chatbots Are Not People: Designed-In Dangers of Human-Like A.I., Public Citizen
• AI Chatbot Says "I'm Human"; Customer Uses a Simple Trick to Test It, Analytics Insight
• AI 'voice cloning' scams are on the rise. Here's how to stay safe, CNN
• 2026 State Chatbot Laws: Key Provisions and Regulatory Trends, Orrick
• Prepare to get manipulated by emotionally expressive chatbots, Wired
• OpenAI Safety Boss Resigns in Latest Executive Departure, PYMNTS