Should you tell customers when they're talking to AI?
By Rodney Noriega, Founder, Universal Enablement · Published Oct 12, 2026 · Updated Oct 12, 2026
Short answer
Yes. People can no longer reliably tell AI from a person in a text chat, so telling them is the only way they'll know. In a 2025 UC San Diego study by Cameron Jones and Benjamin Bergen, an OpenAI model told to act human was picked as the human 73% of the time, more often than the real people. Being upfront protects trust when something goes wrong.
In 1950, Alan Turing proposed a test: if you can't tell a machine from a person by text alone, the machine is acting human. For 75 years that was the benchmark. Now it's been passed. Regulators are responding: in the EU, rules requiring AI systems that talk with people to say they're AI began applying in August 2026. Canada has no general rule like this yet.
That changes things for any business using AI with customers. If they can't tell, they're relying on you to tell them. Customers who find out later often feel misled, even when the help was good.
A practical line to draw:
- Tell them when a customer deals directly with AI: a website chat, an automated phone line, or an AI-written reply sent without review.
- No label needed when a staff member uses ChatGPT, Claude or Copilot to draft, then reviews and sends the message as their own. They own it.
Simple wording works: "You're chatting with our AI assistant. A team member can take over any time. Just ask."
This is one of our five principles: customers know when it's AI.
Related
Start with one workflow. In 30 minutes we'll look at where your team's time goes and tell you honestly whether AI can help.
Book a free 30-minute workflow review