The annual technology, design and entertainment conference known for “ideas worth spreading” used a significant portion of its latest event in Vancouver to try to wrap attendees’ heads around the now fast-moving field of artificial intelligence (AI).
“AI is alarming and amazing all at once,” said TED’s head and curator Chris Anderson during the five-day conference’s opening on Monday at the Vancouver Convention Centre.
Of the nearly 80 people who delivered talks at the conference, at least a fifth directly addressed developments in AI and its potential to uplift life on the planet with medical, environmental or communications breakthroughs — and also disrupt through misinformation, or worse.
TED, which for decades has attracted top names from almost every sector, did the same this year over AI, bringing in key players such as Greg Brockman, co-founder of ChatGPT creator OpenAI, and Tom Graham of Metaphysic.AI, the company that created some famous Tom Cruise deep fakes, as well as several key AI questioners.
‘Smarter and uncaring entity’
Those doubters included Eliezer Yudkowsky, who for 20 years has been studying the rise of AI and the possibility of it overtaking its creators.
Yudkowsky was a last-minute addition to the conference and his short six-minute talk, devoid of any slides or visual elements, earned a standing ovation.
“I expect an actually smarter and uncaring entity will figure out strategies and technologies that can kill us quickly and reliably, and then kill us,” he said.
Yudkowsky has demanded that AI systems be shut down.
But that intervention was not echoed by others at TED, many of whom agreed instead that worldwide co-operation is needed for strong governance to manage AI, which combines computer science and massive data sets to solve problems, and more.
For its part, last summer, the Government of Canada tabled the Artificial Intelligence and Data Act (AIDA) as part of Bill C-27, with the aim of ensuring Canadians can trust the digital technologies they are increasingly using.
“Absent clear standards, it is difficult for consumers to trust the technology and for businesses to demonstrate that they are using it responsibly,” it said.
Despite many TED talkers warning over the need to properly regulate AI, others got involved in the fun and wonder of it.
The conference kicked off with Anderson interacting with Metaphysic.AI’s Graham, whose company is responsible for developing AI that can produce videos of humans often indistinguishable from the real version.
Along with transposing the image and voice of Anderson onto his own face in a live video projected on a screen above them, Graham also released a further video of the technology’s facsimile of actor Tom Cruise expressing his trademark exuberance for being in Vancouver — despite not actually being there.
Graham argues the technology will change the way humans can interact with each other, such as allowing family members to have conversations with grandparents no longer alive.
When Anderson said that might be “creepy,” Graham responded that the human connection would outweigh any initial distaste.
Fake Drake? How to spot AI-generated content
Fake AI-generated videos and images are increasingly available online, but picking them out is becoming harder and harder. Futurist Sinead Bovell takes CBC’s David Common through some of the dupes and shows him how to spot them.
‘Get literate in this technology’
The optimism was continued by Brockman, who did a show-and-tell of the latest versions of ChatGPT and DALL-E, which have been grabbing headlines over being able to converse with humans and to use powerful language processing to draft communications, write code and create images from language.
Brockman had the AI suggest a post-conference meal, put together a shopping list for the items needed to make it, and tweet it out with an image. At the same time he presented the back end of the technology, where he could check how ChatGPT was doing the work.
He defended releasing the technology to be used and improved in real-time by users across the globe, rather than in a closed lab and only greenlit when it cleared any industry or government gatekeeping.
“We’ve got to take each step as we encounter it and I think it’s incredibly important today to get literate in this technology,” he said.
“If ChatGPT wasn’t out there now, then we wouldn’t be having these discussions.”
Safety checks, shortcomings
TED also gave the stage to so-called AI prognosticator Gary Marcus, who called for a non-profit and neutral international governance agency for AI and further safety checks on the technology before it is rolled out.
The University of Washington’s Yejin Choi, meanwhile, pointed out AI’s shortcomings when it comes to common sense and what will be needed (funding) to provide the technology with good data beyond what is freely available online — where hate and falsehoods are plentiful.
Artists at TED spoke about the power AI is giving them to create images, sculptures and music.
“AI can open the mind with remarkable results,” said K Allado-McDowell, a musician and writer who has created books and music in collaboration with AI technology “This will add to the possible forms of expression.”
Ina Fried, a journalist based in San Francisco, who is the chief tech correspondent for Axios and has covered the industry for 25 years, said generative AI has captured the world’s attention through ChatGPT, but the technology is poised to take many more forms — something humans should be optimistic about.
“There’s a lot more that you can do besides typing in a few words and letting the computer have a go at it,” she said.
“You can really have a creative dialogue and that’s why I expect in the next year or two there to be just tons of interesting applications of this major underlying technology.”
But she also agreed with Anderson’s take that AI is one of the biggest issues the conference has ever tried to tackle.
“It’s a major new shift in how humans will react and interact with technology,” Anderson said. “We better get it right.”