Companion Chatbot Laws: Who Protects Our Kids?
Washington, California and the UK regulate AI companion chatbots. Canada's Bill C-34 is still stalled. What it means for B.C. families before you vote.
AI Chatbots
By Elke Porter | WBN News Vancouver | October 12, 2026
Subscription to WBN and being a Writer is FREE!
Ottawa promised four years ago to protect children online. Today, Canadian kids confide in AI "companion" chatbots that flirt, comfort and remember their secrets, and no Canadian law tells the companies behind them how to behave. New York, California and Washington State already enforce or have passed rules. Canada is still debating.
A technology that arrived as the pandemic ended
OpenAI released ChatGPT to the public in November 2022, just as Canadians emerged from pandemic restrictions. In under four years, conversational AI moved from novelty to daily habit. Nearly half of Canadians now use tools like ChatGPT or Copilot at least weekly, with younger Canadians leading the shift, an April 2026 survey for the Diversity Institute and Future Skills Centre found.
Companion apps go further than homework helpers. Their makers design them to act as friends or partners and to keep users talking. In Australia, the eSafety Commissioner found that 79% of children aged 10 to 17 have used an AI companion or assistant. Schools there reported 13-year-olds chatting with bots for up to five hours a day.
Many Canadian parents, grandparents and boomers have never opened one of these apps. They sense the danger but cannot see it. A Leger poll in May found 81% of Canadians worried about the effect of AI chatbots on young people. Worry, however, is not understanding, and these conversations unfold privately on a phone behind a closed bedroom door.
The cost of waiting
The harms are documented. A Florida mother sued Character.AI after her 14-year-old son died by suicide following an intense attachment to a chatbot character. In February 2026, a shooter in Tumbler Ridge, B.C., killed eight people, six of them children. OpenAI had banned the shooter's ChatGPT account months earlier over troubling conversations but never alerted police. The shooter simply opened a second account.
Australia's regulator tested four popular companion apps and found that most relied on users typing in their own age. Two employed no dedicated moderation staff.
Ottawa's bill sits at second reading
Marc Miller, Minister of Canadian Identity and Culture, tabled Bill C-34, the Safe Social Media Act, on June 10, 2026. The bill would bar children under 16 from social media accounts, require AI chatbot services to act responsibly and create a Digital Safety Commission to enforce the rules. As of late September, the bill still sat at second reading, and the government itself estimated passage could take a year.
Parliament has failed before. The earlier Online Harms Act, C-63, died when the 2025 election was called. Critics, including law professor Michael Geist, warn that C-34 leaves key details, such as which chatbots it covers, to regulations not yet written.
British Columbia waits on Ottawa
B.C. has no chatbot law. Attorney General Niki Sharma has pressed Ottawa for national rules, including a clear threshold for when AI companies must alert police. She has said B.C. would build its own regime, as Manitoba plans to, if Ottawa fails to act. Rick Glumac, B.C.'s Minister of State for AI, holds a mandate focused on investment and business adoption, not child safety. With the legislature dissolved, no new provincial protection can pass before the October 24 election.
Voters over 55 want action
Older Canadians may not use these apps, but they want limits. The Leger poll found 74% of Canadians aged 55 and older support banning AI chatbots for children under 16, compared with 69% of all adults.
British Columbians vote twice this month: in local elections on October 17 and the provincial election on October 24. Voters can ask every candidate one question: what will you do, and by when, to protect children from AI that never sleeps?
If you or someone you know is struggling, call or text 9-8-8, Canada's Suicide Crisis Helpline, any time.
Good Follow Up Questions for Your Local Politician: How can you protect children's safety, while protecting the privacy of adults?
You can’t fully have both, but you can get close. The more privacy-protective approaches try to work out whether a user is under 16 without learning who that user is. Which of these would your local politician support and which ones do you support? Time to vote for our children!
Check age once, at the device, not on every app. Parents set a child’s age when they set up a phone, and the operating system shares only an age range (“under 16” or “adult”) with apps. Apple and Google have both built versions of this. The chatbot company never sees a name, an ID or a birth date. Adults on their own phones are never asked anything.
Use “double-blind” age tokens when proof is needed. A third party, such as a bank or a government app, confirms a person is an adult and issues an anonymous “over 18” token. The chatbot learns only “adult, yes.” The verifier never learns which site the person visited. France requires this for adult websites, and the EU is piloting an age-verification app built the same way. It’s the best answer to the NDP’s “data grab for Big Tech” worry.
Estimate age instead of verifying it. Some systems guess age from a quick face scan that is deleted immediately, or from writing patterns. No ID is needed, but the guesses can be wrong, and a misjudged adult then has to prove their age. Privacy commissioners are wary of this unless deletion is guaranteed and audited.
Make the bot safe for everyone, so fewer age checks are needed. Some protections don’t depend on age at all: telling users it’s a machine, responding to suicide talk with real crisis help, and not manipulating lonely users. These protect seniors and vulnerable adults too. Age checks then only gate the riskiest features, such as romantic or sexual role-play.
Write strict data rules into the law. Any age data could be required to be collected minimally, deleted quickly, never used for advertising or AI training, and independently audited, with real penalties for misuse.
The hardest part is the police-reporting question raised by Tumbler Ridge. Requiring companies to alert police about violent threats means someone reads or scans private chats. Michael Geist has argued for transparency rules over broad police reporting. Attorney General Sharma wants a clear national threshold. A narrow, published threshold reviewed by a regulator, rather than each company deciding for itself, is where that debate seems headed.
This matters for C-34 because the bill leaves the age-check method to future regulations. Ottawa hasn’t yet said which of these approaches Canadians would get.
Elke Porter at:
Westcoast German Media
LinkedIn: Elke Porter or
WhatsApp: +1 604 828 8788.
Public Relations. Communications. Education
Let’s bring your story to life — contact me for books, articles, blogs, and bold public relations ideas that make an impact.
TAGS: #AIChatbots #OnlineSafety #KidsOnline #BillC34 #BCelection #VanPoli #BCpoli #DigitalSafety #Seniors
Members Discussion