A group of advocates rallied at the Vancouver Art Gallery on Aug. 26 to call for clear ethics and safety standards in the artificial intelligence (AI) industry.
Matias Raski recently founded ALIGN the World, a non-profit organization on AI safety, to raise awareness about the “big risks” associated with AI.
“I wish people understood how rapidly the technology is developing, how likely we are to be surprised by new capabilities that emerge through AI systems that are trained on larger datasets with more computing power, and just how difficult it is to predict what the systems are likely to be able to do,” Raski says.
Raski is also concerned about AI potentially becoming something people can not control. Himself and seven other advocates with the non-profit handed out pamphlets addressing their concerns with AI. Each pamphlet included a QR code to an online petition Raski started. It calls on the Government of Canada “to guide the development of AI to protect Canadians.”
Khue-Tu Nguyen, who helped create ALIGN the World, also attended the rally.
“I think it’s a really important topic that although a lot of academics are talking about it, a lot of scientists are talking about it, [as well as] a lot of people in the AI industry, I just don’t think this conversation has permeated the general population yet,” Nguyen says.
“It’s such a big issue that will impact all of us. I think that’s why we’re here — just to talk to everyday people in downtown Vancouver to spread the word.”
While Innovation, Science and Economic Development Canada’s Bill C-27 is under consideration, which includes the Artificial Intelligence and Data Act to regulate AI, Raski says he would like to see more action, specifically on addressing potential “larger-scale safety issues.”
“We’re urging the Canadian government … to invest in AI ethics, safety, and governance research and to take part in this global conversation about coordinating to prevent harms from AI,” he says.
Currently, Canada does not have any regulatory framework that specifically addresses AI.
Both Nguyen and Raski, who graduated from the University of British Columbia’s medical school, agree that like new medicines and therapies, AI systems should go through safety tests before they are released to the public.
“It’s a no-brainer, and I don’t think anyone can argue that we don’t need [safety testing] in medicine. We definitely need it,” Nguyen says.
“In the same vein, … it should also happen in AI. It’s such a new technology that can affect so many people in very rapid ways as well, from an environmental, healthcare, information, [and] media perspective. I think if it has such wide-ranging reach and effects, it should be subject to similar ethical and safety regulations.”
ALIGN the World and its “align AI with humanity” slogan are named after AI alignment, an industry term that means ensuring AI systems behave in the way humans want.
Despite having safety concerns, Raski says his non-profit is not anti-AI.
“I do want to emphasize that I don’t think any kind of catastrophic outcome with AI is at all inevitable. I do think the benefits are likely to be huge, both in terms of scientific breakthroughs and economic gains, [regarding] doing dangerous jobs that we probably shouldn’t want humans to have to do and various other things,” he says.
“I think realizing those benefits and preventing those harms [needs] to be a collaborative effort between AI companies and governments.”
ALIGN the World will have another rally on Oct. 21 on Parliament Hill in Ottawa.
For more information, visit www.aligntheworld.com/.