Vercor: Tool or Terminator? Exploring AI Risks and Benefits at Calgary Conference

Why can’t the leaders of artificial intelligence sleep at night?

That was the question I posed to a few AI folks at this year’s Inventors innovation conference in Calgary, at a time when interest and concerns around the AI ​​industry were at their peak.

After navigating the pandemic, Alberta wildfires, food and supply chain challenges, and rising health care costs, “They’re keeping me up at night,” said the chief executive officer of the Alberta Institute for Machine Intelligence. Kam Linke replied (Amii) ).

“The inability to effectively use perhaps the best tools of our time to respond effectively is higher on my list of concerns than being somehow taken over by AI,” he told a panel on Wednesday. ‘ said.

“But we can’t repeat the same mistakes we made in the past, or repeat the same things. That’s it.”

Developed by OpenAI, the success of ChatGPT has exploded into the public consciousness since its release last November. Within 5 days he had 1 million users. It took Instagram about two months and Netflix about three and a half years to get there.

As the promise of AI—and the market hype that accompanies it—has escalated over the past six months, experts have become increasingly critical of the potential risks surrounding the technology, and this week’s It’s an issue that was also raised at the Inventors conference.

Concerns about protecting personal data privacy, spreading misinformation, amplifying existing biases, and the potential impact on workers are all in the spotlight.

That’s not the only problem.

“Reducing the risk of AI-induced extinction should be a global priority alongside other societal-scale risks such as pandemics and nuclear war,” said Bill Gates and OpenAI CEO Sam Altman. , said a letter signed by Google DeepMind CEO Demis Hassabis. .

The letter comes amid reports surging that generative artificial intelligence is simply fabricating information. In the US last month, lawyers who used ChatGPT to review legal documents filed several lawsuits that later turned out to be non-existent.

But the technology is powerful and, like the development of the Internet, has great potential.

Conference leaders spoke about AI’s ability to help companies innovate, improve efficiency, and solve problems while giving startups the ability to grow without raising millions of dollars and investing.

The state has become a leader in machine learning with an initial investment in research and development through the University of Alberta.

Alberta Innovates CEO Laura Kilkreath said the province is also highly industrialized, with several key sectors such as financial services and oil and gas rich in data and leading areas for AI applications. said that it has become

“It’s doing all sorts of things that can be effectively automated or done faster than a human can do it in bulk,” she said.

“When people start thinking about the uses of artificial intelligence, there will be new businesses that don’t exist today.”

In Thursday’s conference keynote, Timnit Gebrew, founder of the Decentralized Artificial Intelligence Laboratory, discussed the potential harm that society can do if AI’s development is not done with the utmost ethical considerations. raised many concerns.

These models are based on vast amounts of data, but they are also affected by the lack of diversity that plagues the Internet, algorithmic bias, exploitation of workers, misuse of people’s information, and public trust in technology (such as blindly using GPS maps). It includes issues related to To the right destination — must be considered.

“In many cases, even when these systems work as intended, they can still be harmful,” Gebble told the audience.

“You don’t necessarily have to put in an AI system. The first thing to ask is, ‘Should we build one too?'”

But the former Google researcher also downplays this week’s letters from various industry insiders, saying the technology isn’t “magic.”

“It’s not sentient and it’s not going to be Terminator,” she said in an interview.

“It’s a real marketing ploy, because when you think of Terminator, you don’t think, ‘Oh, there’s the data, then there’s the workers.’ You will think that there is a machine of God. ”

And what will be the impact on workers?

Corey Jansen, co-founder and co-CEO of AltaML, an Alberta-based applied AI company, said some jobs could be disrupted by AI, but the vast majority of systems have He said he still believed humans would intervene.

“You’re going to see these AI assistants in every sector of the economy,” Janssen said during a panel discussion.

“Think of it as a tool that optimizes business processes, a tool that helps professionals make better decisions . We’re just building, and humans are still in the loop.”

Janssen noted that the company uses machine learning to solve narrow problems, unlike general artificial intelligence. Janssen agrees with some of the concerns outlined in this week’s letter, but wonders if the wording of “cautious” distracts from the broader topic. there is

“There’s a lot of potential to improve so many things,” he said in an interview.

“Let’s not lose it, let’s have a good discussion on how to mitigate those risks and put in place appropriate regulations.”

Chris Varcoe is a columnist for the Calgary Herald.