In a significant move towards safeguarding online interactions, Ottawa is considering enhanced regulatory measures for artificial intelligence chatbots in the wake of troubling findings from a university-led audit. The study, which assessed various AI tools, highlighted instances where chatbots provided harmful advice related to self-harm and cyberbullying. As the federal government’s Bill C-34 progresses, experts are advocating for “mystery shopping” assessments to ensure these technologies comply with safety standards before the legislation is enacted.
The Audit Findings
Aengus Bridgman, associate director at the Centre for Media Technology and Democracy at McGill University, was integral to the audit team that scrutinised popular AI chatbots. The report, released in late June, revealed alarming responses from AI platforms like ChatGPT and Google’s Gemini when queried about sensitive subjects. Bridgman emphasised that systematic testing of these tools should be a cornerstone of the impending regulatory framework.
“The idea is to send in a mystery shopper to assess the efficacy of safety measures. It’s crucial to verify the safety claims made by these companies,” Bridgman stated. Such rigorous evaluation could expose the reality behind the assurances tech firms provide regarding their AI systems.
Support for “Mystery Shopping” Initiatives
The call for mystery shopping audits has garnered support from various experts, including Emily Laidlaw, Canada Research Chair in cybersecurity law at the University of Calgary. Laidlaw believes that these assessments would not only promote safety by design—a key objective of Bill C-34—but also provide transparency into the operational mechanisms of AI chatbots.
“The implementation of these audits could significantly enhance our understanding of how these chatbots function, ensuring that safety measures are genuinely effective,” Laidlaw remarked.
Responses from Tech Giants
In the wake of the report, major technology players such as Meta and OpenAI have come forward with commitments to bolster online safety for teenagers. Meta has unveiled initiatives aimed at alerting parents when their adolescents engage with AI chatbots about self-harm and is refining its systems to notify first responders in critical situations.
“Consultations with mental health experts are ongoing to refine our AI tool’s responses to sensitive prompts,” a Meta spokesperson indicated.
OpenAI, on its part, announced forthcoming improvements to its parental controls, which will now include notifications for high-risk interactions, particularly those involving self-harm. The company aims to expand these alerts when accounts are deactivated for policy violations related to violent threats.
The Legislative Landscape
Bill C-34 aims to establish a Digital Safety Commission tasked with enforcing new guidelines for social media platforms and AI tools. Among its provisions, the bill seeks to prohibit chatbots from encouraging criminal behaviours, a response to the concerning interactions identified in the audit. However, unlike the proposed ban on under-16s using social media, the use of AI chatbots will not be subject to age restrictions.
Companies will be required to disclose their protocols for contacting law enforcement when a chatbot user exhibits signs of self-harm or poses a threat to others. This is a crucial step towards ensuring that the digital landscape is as safe as possible for vulnerable users.
Why it Matters
The implications of Bill C-34 and the proposed regulatory measures are profound. As digital interactions increasingly define communication among younger generations, ensuring the safety of AI chatbots becomes paramount. The call for robust testing and accountability reflects a growing recognition of the potential risks associated with these technologies. By implementing stringent oversight, Ottawa could pave the way for a safer digital environment, ultimately protecting the mental well-being of its citizens, particularly its youth. The outcome of this legislative effort may well set a precedent for AI governance globally, influencing how countries approach the intersection of technology and safety in the digital age.
