Dr. Mary Wells, the Dean of Waterloo Engineering, urged the Canadian government in a memo not to allow anyone under 18 years of age to interact with artificial intelligence (AI) companion models “that attempt to develop an emotional bond or have been designed to be manipulative or addictive in nature.”
Wells submitted that recommendation and others after being appointed last year to an AI strategy task force by Federal AI Minister Evan Solomon, who is expected to publish a new government AI strategy as early as next month based in part on input from the expert panel.
In her memo, Wells recommended a tiered approach to categorizing AI risk, focusing on effects on people rather than the underlying technology.
She also urged prohibiting the use and deployment of AI systems that deliberately deceive users into believing they are interacting with a human and said AI-generated content should be clearly identified.
“Chatbots must be required to identify themselves as such,” Wells wrote, according to a story in The Globe and Mail this week. “AI-generated media should require visible watermarks and/or other identifiers.”
One of 26 members on the task force, Wells also said people should have the right to know if their data has been used to train a commercial AI system, or is being accessed by an AI system, and be able to withdraw permission.