It's the year 2050 and artificial intelligence is no longer the newest topic but is integrated seamlessly in our everyday lives.
At least that's what experts believe will be the case.
The technology has made leaps in the last year as open-sourced intelligence like ChatGPT and Google Bard allow the public a glimpse into what the future could hold.
AI has been touted as a way to solve many of the world's issues, including in health care, transportation and infrastructure, but at the same time, questions are raised about its efficacy.
From avatars giving mental health advice to generated internet photos, AI has quickly found its way into the lives of Canadians, but its future is still unknown.
CTVNews.ca asked three experts what AI will look like in 2050 including what it will impact most and what Canada needs to do to be ready.
TODAY'S AI 'TEND TO HALLUCINATE'
"It's actually incredibly difficult to predict so much time ahead," Vered Shwartz, assistant professor at the University of British Columbia said.
To understand what the future could hold, Shwartz, also a Canadian Institute for Advanced Research (CIFAR) AI chair, looked to the past.
"In 1996, the internet was new, you had to connect to the cable, Amazon started as an online bookstore and no Google, video, social media and smartphones," she told CTVNews.ca in an interview. "If you just think about how much changed since then, and the fact that things are progressing faster now, it's actually very hard to predict."
But she says with an understanding of how AI is advancing now, it is possible to see how the world will become more automated.
"(It will be) even more ubiquitous in our lives and I think it will get better," she said. "There are a lot of limitations right now but I think many of them will be solved. I think it will enhance a lot of jobs and likely also replace them and create some too."
Jobs heavy with admin tasks and paperwork would likely benefit from AI, Shwartz said, including law. She can see AI aiding lawyers by automating "mundane tasks" like writing contracts, but says this is still a work in progress.
"The main problem today is with generative AI tools…They're just not reliable because they tend to hallucinate and make up details, which can be very problematic in that context," she said.
Recently two U.S. lawyers blamed ChatGPT for adding fictitious details to their legal research. They could face punishment from the judge over filling in a lawsuit that included fake case references.
AI could also help people learn new languages easier, Shwartz said. For French and English, AI can amplify learning due to its large pool of resources on the internet.
"But if you look at an Indigenous language that doesn't have a lot of written text on the web, so standard solutions don't actually work well for such low-resource languages," she said. "So you have to do more research and come up with different solutions to actually make that work, but AI can definitely help with that."
Shwartz hopes AI will also be used specifically to optimize public transportation.
"I don't think there will be flying cars. I don't necessarily think that's an efficient solution. I think self-driving cars…will become more mature," she said. "I really hope that in 2050 we will not still be standing in traffic."
THE FUTURE OF CUSTOMER SERVICE
While Shwartz envisions a world with more automation, Jackie Cheung, a McGill University professor and consultant with Microsoft Research, imagines more avatars assisting humans.
"There are already many start-ups and companies working on that type of (avatar) application with kind of generative AI technology," Cheung, also a CIFAR AI chair, told CTVNews.ca in an interview.
He believes people will be interacting with avatars at grocery stores, banks and restaurants in the "not too distant future."
"I think full automation in the short term is probably a mistake in terms of consumer experience and how it affects their day-to-day lives," Cheung said. "As systems improved, there are some situations where it's probably better than the current state."
One area Cheung could see AI benefiting is within the airline industry, where customer service could be given through technology, but he says AI avatars could spark an emotional response in humans that we may not account for.
"It could cause a lot of emotional distress with the people who interact with these systems and develop bonds," he said. "There are lots of longer-term consequences that we haven't really thought much about."
Some areas of our lives have already been infiltrated by AI, Cheung said, like search queries on the internet that pop up as options.
The question-based way of searching and receiving specific examples is one way Cheung believes AI will aid in customer service in the future.
"We won't immediately start using them like at full capacity," he said. "It will take time to figure out where they're the most useful and where maybe they're less useful."
DR. AI
Searching for research could also be done by an AI personal assistant, Rahul Krishnan, a CIFAR AI chair, told CTVNews.ca in an interview.
"Once we have these large-scale multimodal models that can be personalized things, that's where we get to do very interactive chatbots that can assist us in many open-ended ways," Krishnan, a professor at the University of Toronto and affiliated with the Vector Insitute for Artificial Intelligence, said.
Krishnan imagines wanting to go on a hiking trip and asking his own personalized assistant to provide a plan for what he might need.
In his mind the assistant could also order groceries with approval, remind about appointments and schedule meetings.
"This is my imagination in the year 2050, but I do suspect it may come a lot sooner than that," he said.
Krishnan focuses on AI machine learning, which is how the technology learns and then implements ideas. One way he believes AI could change Canada is through health care.
"It's going to change health care because I think health care is just a very big heterogeneous system," he said. "Small, incremental improvements over time is really going to transform it from what we have today to something where clinicians are able to spend a lot more time with their patients than they currently do in front of a computer."
In January, something Krishnan believes could be given to AI eventually.
The technology could help physicians with taking detailed patient notes or summarizing lengthy clinical history.
AI could also assist doctors in finding issues with patients, Krishnan said.
"You can imagine a pathologist looking at a very, very large image, trying to identify abnormalities in cells," he said. "What might take several hours to do might be shortened to a 30-minute period because it's not just a single person working on this image, it's a single person working with an AI-assisted device that has been trained on millions and millions of other images."
The tech comes with solutions to aid humans in ways many experts cannot predict would be possible today, but Krishnan says AI is a step forward.
"We as a society choose to explore to ensure that the future continues to evolve in a way that's beneficial to everyone," he said.
