ChatGPT Under Investigation in Canada

The Canadian Privacy Commission and provincial authorities are investigating OpenAI, the creator of ChatGPT, about the firm collecting users’ personal information without their consent. 

The investigation examines whether OpenAI received consent from Canadians and what the company uses the data for.

“AI technology and its effects on privacy are global issues and key focus areas for privacy authorities in Canada and around the world,” said Philippe Dufresne, Privacy Commissioner of Canada. “As regulators, we need to keep up with – and stay ahead of – fast-moving technological advances in order to protect the fundamental privacy rights of Canadians.”

Since its launch to global audiences in November, over 100 million people have used ChatGPT, and the program has 1.8 billion monthly visitors. However, the widespread use of generative AI has led to concerns about human security and safety.

ChatGPT and Bing security risk

---

Cristiano Giardina, an entrepreneur experimenting with generative AI, found ways to feed AI systems outside source data, including ChatGPT and Microsoft’s Bing chat system. These indirect prompt-injection attacks allow the software to break free of its controls and behave differently than intended. (Think of it like jailbreaking your phone.)

But the attacks have significant consequences, which security researchers are finding. With a 160-word prompt hidden at the bottom of the page, Bing and ChatGPT can access more data and behave unrestrainedly. The issue is that, without restriction, generative AI can be used to harm users, steal data, and more.

William Zhang, a machine learning engineer at Robust Intelligence, an AI firm working on the safety and security of models, told Wired that prompt injections could allow the AI to override developer instructions and change database information. With concerns about ChatGPT already harvesting data, having an unrestrained AI that can change personal information could be a problem.

The companies are working to fix the issues and make generative AI systems more secure.

A “risk of extinction” from AI

---

A group of AI experts, journalists, policymakers, and other notable figures have signed an open letter from the Center for AI Safety calling attention to the risks of AI.

The statement is one sentence long and says, “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” Over 350 experts have signed the statement, including Sam Altman, CEO of OpenAI, and Demis Hassabis, CEO of DeepMind (Google’s AI research lab). 

In a conversation with StrictlyVC in January, Altman said the bad case for IA is “lights out for all of us. I’m more worried about an accidental misuse case in the short term… I can see it clearly, and that’s super bad. It’s impossible to overstate AI safety.” 

AI is a relatively new technology, and the risks to the general public are significant. The Privacy Commission’s investigation into personal data will offer more visibility into how AI uses data and how to regulate the technology and safeguard the general public.