ChatGPT a 'landmark event' for AI, but what does it mean for the future of human labour and disinformation?
Chatbot is the newest artificial intelligence tool on the block — and its responses feel scarily human
ChatGPT is the newest artificial intelligence tool on the block, and it can do anything from create poetry, write an intro for a radio show — like The Current — and even help with homework.
"I introduced this program to the class last Tuesday," said Ethan Mollick, a professor at the Wharton School of the University of Pennsylvania.
"People were asking it to find the error in their computer assignment, polishing up drafts, explaining a concept to them like they were five years old," he told The Current's Matt Galloway.
ChatGPT is an interactive program trained by AI research lab OpenAI. It was launched on Nov. 30, but has already amassed more than a million users, according to OpenAI CEO Sam Altman.
It's based on OpenAI's GPT-3.5 language model, which is an improved version of GPT-3 language model, released in 2020. Users can type a question or prompt into the program, and ChatGPT answers with a response designed to mimic that of a human — which the program does well, according to Mollick.
"This is one of several of these models … and [it] feels much more realistic and has really crossed a line where the kind of work it's doing actually feels human as opposed to talking to a machine," he said.
Michael Wooldridge, a computer science professor at the University of Oxford, says it's a "landmark event" for AI.
"These tools can produce language, which is certainly at the level of an undergraduate student or a typical office worker," he told Galloway. "That feels like a really important moment."
Its success has also raised some questions about whether it can replace search engines like Google.
"Right now, it has a tendency to lie with extreme competence, or hallucinate and make up facts, and so if you use Google, you'll be disappointed," Mollick said. "It's also not connected to the internet, so it's not going to be able to give you the answers to your questions the way you might have wanted it to."
"But it's not just Google, right? The question is also if it can create lectures and syllabi and writing for you, what else does that do that Google doesn't already do right now?"
For the vast majority of people, it's just going to be another tool that they use and it's going to make them more productive.-Michael Wooldridge, computer science professor
Replacing human jobs?
The writing of lectures and syllabi — all at the ease of a simple request — also raises questions about whether AI programs like ChatGPT can replace entire jobs and professions.
Mollick said the thought is both liberating and scary "because I don't think we know quite what the expectations would be about what's done by humans … and what kinds of jobs it should replace."
Wooldridge says the automation of labour is far from new. Since the first time somebody hooked a plough to an ox, automation of labour has relieved human beings of a job, but it didn't eliminate the need for human labour.
"For the vast majority of people, it's just going to be another tool that they use and it's going to make them more productive," he said. "It's going to take the strain off a tedious and difficult task and make them better at that."
But programs like ChatGPT can affect how a variety of different jobs do work, according to Mollick, and that raises some tough questions to ponder.
"Do I think most programmers will be out of a job? No," he said. "But do I think that some set of programmers might be able to be replaced by AI in a better way? Possibly."
"I don't think it's 100 per cent clear that it's going to have the same effect on every industry and every job, and I think the fact that it will affect so many at once is what makes this a little bit unusual."
The role of social media
For Wooldridge, one of the more pressing concerns with the emergence of programs like ChatGPT is the spread of AI-generated fake news.
"One of the challenges is … you can create the skeleton of a fake news story, ask it to produce 100 different variations on that, and then just use 100 different fake Twitter IDs or Facebook IDs to start spreading that," he said.
I don't have the ability to browse the internet or generate news, and I don't have the ability to prevent the spread of fake news.-ChatGPT's AI
This isn't helped by the trust issues social media platforms like Twitter are currently experiencing, says Wooldridge.
"The issue of verified IDs on Twitter, I thought it was a really big one," he said. "I mean, the blue tick that you used to get on Twitter meant that you could have some confidence that you were looking at a real person — and it's not clear that that's true anymore."
One way this issue could be fought is by inserting a hidden digital signature into text generated by an AI program in order to signify it was produced by a computer and not a human journalist.
For now, the onus might be on social media platforms to do a better job filtering out fake news to prevent AI-generated news from spreading — and according to the AI behind ChatGPT, the user to distinguish between fake and real news.
"I don't have the ability to browse the internet or generate news, and I don't have the ability to prevent the spread of fake news," the AI said, in part, to a question from CBC Radio about ChatGPT's role in preventing the spread of fake news.
"It is important for users to be able to distinguish between real and fake news, and to verify the accuracy of any information they encounter before sharing it. This is especially important when it comes to AI-generated content, as it can sometimes be difficult to distinguish from real news," ChatGPT said.
Produced by Benjamin Jamieson and Paul MacInnis.
