Ottawa is asking whether people should know when they are talking to an AI. One example it chose is close to home for any business that texts its leads.
What happened?
Innovation, Science and Economic Development Canada (ISED) ran a consultation on AI transparency from July 23 to September 23, 2026. One of its five areas is helping people know when they are interacting with an AI system. ISED says input will be summarized in a What We Heard report. On October 23, 2026, the Office of the Privacy Commissioner of Canada published its submission, agreeing that transparency issues can arise in all five areas: AI-generated content, AI interaction, information about AI systems, AI incidents and AI agents.
ISED’s discussion paper includes a case study. In May 2026, a Toronto man asked a dealership about selling his car back. The dealership started a text conversation under the name “Quinn”, which offered a price, said the dealership would consider his counteroffer and set a meeting. A sales consultant then revoked the offer and explained that Quinn was an AI chatbot that had made it in error. Quinn never said it wasn’t human. The dealership honoured the original offer after the man took his story to the media. The same paper notes that proposed Bill C-34 includes rules against chatbots deceptively posing as a human.
What it means for your business
There is no new disclosure rule from this consultation yet. Treat it as direction, not law. But the case study shows the cost isn’t only regulatory. An undisclosed bot made a commitment the business didn’t want to keep, and the business paid for it in public.
AI texting and voice are often the right tools for revenue follow-up, the step where many owner-led firms lose leads. SMS follow-up and reactivation works because it is fast and it doesn’t forget. The answer isn’t to slow it down. It is two design rules:
→ say it’s AI in the first message, in plain words → cap what it can commit to (prices, discounts, dates) and send anything outside the cap to a person
Then log every conversation: what the bot said and what a person changed. When a bot oversteps, you find it in your log, not in the news.
This is the pattern to expect. Once AI handles the follow-up itself, the limit moves to trust and to checking the work. That is where owner attention should go next.
What to do this week
- Read the first message each of your bots sends. If it doesn’t say it is an AI, change it.
- List what each bot may offer. Remove prices and commitments it shouldn’t make alone, and route those to a person.
- Check that your AI voice agents say they are AI in the first sentence of the call.
What to watch
ISED’s What We Heard report, and whether it proposes a disclosure rule for business chatbots. Also Bill C-34’s progress in Parliament.