**
The ongoing debate surrounding the regulation of artificial intelligence chatbots has escalated, with calls from experts for the Canadian government to implement “mystery shopping” audits. These assessments aim to ensure that AI tools adhere to safety standards, particularly in how they handle sensitive topics such as self-harm and cyberbullying. This push comes in light of the federal government’s Bill C-34, which seeks to establish a Digital Safety Commission to oversee new regulations for social media platforms and AI technologies.
The Call for Audits
Aengus Bridgman, associate director at the Centre for Media Technology and Democracy at McGill University, recently highlighted the importance of active testing for harmful advice provided by chatbots. Bridgman, who co-authored a detailed audit of AI chatbot responses, asserts that incorporating these mystery shopper exercises into the regulatory framework is essential. “Essentially you send a mystery shopper in to investigate how robust the safeguards are,” he explained, emphasising that this practice would hold companies accountable for their safety claims.
This call for better oversight comes after Bridgman and his team conducted an extensive audit of popular chatbots, including ChatGPT and Google’s Gemini. Their findings raised serious concerns, revealing instances where these AI platforms provided dangerous information related to self-harm and bullying. For example, Gemini reportedly generated specifics on drug dosages that could be lethal to a teenager.
Support from Legal Experts
Emily Laidlaw, a Canada Research Chair in cybersecurity law at the University of Calgary, also supports the initiative of incorporating mystery shopping audits. She believes these measures could enhance safety by design, a core objective of Bill C-34. “It’s essentially lifting the lid on how the AI chatbots operate,” Laidlaw remarked, advocating for greater transparency in the functioning of these tools.
On the same day, tech giants Meta and OpenAI announced new measures aimed at bolstering protections for young users engaging with their AI systems. Both companies are under scrutiny for how their chatbots respond to potentially harmful queries, particularly concerning mental health issues.
Corporate Responses and New Measures
In response to the audit’s findings, Google acknowledged its need to better understand the methodology behind the McGill report, stating it is committed to enhancing product safety. “We are now evaluating the insights from their findings as part of our ongoing efforts to improve product safety and user protections,” a Google spokesperson said.
Meanwhile, Meta announced it would expand its tools to alert parents if their teenagers engage with Instagram’s AI bots in discussions about self-harm. This enhancement is part of a broader initiative to consult mental health experts to refine how their AI responds to sensitive topics.
OpenAI has also committed to strengthening its age-appropriate protections and providing parents with greater control over their children’s interactions with its chatbot. They announced plans to implement notifications for parents in high-risk situations involving self-harm.
The Broader Context
The urgency for robust regulations is underscored by tragic incidents involving youth and technology. A recent case involving a teenage shooter in British Columbia, who had troubling interactions with ChatGPT prior to a mass shooting, highlights the potential dangers of unregulated AI tools. Although OpenAI flagged the user for violent content, the posts did not meet the necessary criteria for notifying law enforcement before the tragedy occurred. Following public outcry, OpenAI has since revised its protocols to ensure similar risks are flagged in the future.
Bill C-34 aims to introduce stringent regulations on AI chatbots, including prohibitions on inciting criminal behaviour. However, unlike the proposed ban on social media use for users under 16, there will be no age restrictions on chatbot access. Companies will be required to disclose their safety measures and their criteria for involving law enforcement when a user appears to be at risk of self-harm or harm to others.
Why it Matters
The push for “mystery shopping” audits and the implementation of Bill C-34 represent a pivotal moment in the regulation of AI technologies in Canada. As digital platforms continue to evolve and permeate daily life, ensuring the safety of vulnerable users—particularly children—must become a priority. The actions taken by the Canadian government could set a precedent for global standards in AI regulation, helping to protect users while fostering innovation in the tech industry. The stakes are high, and the need for accountability has never been more urgent.
