Artificial intelligence can now write for itself

Artificial intelligence can now write for itself

While many people may not think about it, artificial intelligence (AI) is a constant part of our lives, being used in apps to tailor our food choices and music tastes.

However, the latest AI tool is transforming the very words we read on a page or computer screen, by automatically generating passages of text on any topic.

“I call this the big bang moment for artificial intelligence,” says technology anthropologist Giles Crouch.

Crouch is talking about ChatGPT, the newest AI assistant. Give the bot an assignment, and it can translate what it has learned from reading the internet into news articles, essays, poetry, and music lyrics.

Crouch says the technology has both benefits and risks.

“The big challenge for search engines is people writing lots of content for marketing and they’re all trying to get ranked up at the top of the search engine,” he says.

Skewing search engines is just one pitfall, says Crouch. The technology also brings a myriad of ethical considerations to the forefront around copyright, plagiarism, and misinformation.

“We’re already seeing too that ChatGPT and other AI tools are creating false facts,” says Crouch. “So, think about this in terms of conspiracy theorists and we see out there with far right and far left groups. Now they’re going to use these facts as if they’re actually true, they’re going to use that content and then create videos [and] articles.”

Issues Crouch says society as a whole is going to have to deal with.

“I still think the true creativity comes from humans,” says Mark Hobbs, who has been working with AI for the past decade.

His firm, Fundmetric, uses AI as a tool to help non-profits generate more funds, turning vast amounts of data into usable information that can help target both current and new donors.

He says one of the keys to using AI is knowing the biases that can lie behind the code.

“It’s a constant vigilance, that’s required, to be thinking about what voices aren’t being brought into the datasets, what isn’t being considered,” says Hobbs. “And then going the extra mile and saying, ‘How do we address those issues.’”

The lead data scientist for Fundmetric, Greg Lee, says he’s not worried that AI such as ChatGPT will take over jobs.

“AI – at least right now – is a tool that people can use,” he says. “There’s an incremental development all the time that leads to these breakthroughs, but when that will turn into something scary. That doesn’t seem soon to me based on my experience.”

Many apps using the ChatGPT technology target online content creators such as bloggers and marketers.

However, travel and food blogger Cailin O’Neil of NovaScotiaExplorer.com says she’s not considering using AI to write her online content so far. 

“Because to me, it’s just not authentic,” she says. “I’m creating content because I enjoy it and I love it, and taking that away and giving it to AI is taking away the fun part of my job.”

She’s also concerned about other people passing off AI content as their own.

“A big worry is plagiarism,” O’Neil adds. “And not giving the proper information as well, because you’re just telling this computer system to ‘Go out and find me the best 10 cheeseburgers in Toronto,’ and you’re not really experiencing it for yourself.”

Crouch’s prediction is for a “very messy” 10 to 15 years for the space as society tries to figure out how best to utilize the technology, all things considered.

EXTRA: CTV Atlantic’s Heidi Petracek asked ChatGPT to “write a poem about CTV News at Five,” and this is what it created.

A poem about CTV News at Five written by AI assistant ChatGPT.

Source link

Share This
COMMENTS

Leave a Reply

Your email address will not be published.