Computer programmers and financial workers are just two kinds of white-collar jobs that may have to worry about artificial intelligence and chatbots like OpenAI's ChatGPT.
Tech jobs like programmers could be at risk. ChatGPT can write and debug some code, for instance.
While AI can help those in tech save time on some coding work, Mark Muro, a senior fellow at the Brookings Institution, told Insider that with things like ChatGPT being able to quickly create code, it may mean not as many workers are needed.
According to Bryan Seegmiller, assistant professor of finance at Kellogg School of Management, per a recent newsletter from The Insightful Leader, jobs that include data analysis could also be at risk. In particular, "those who specialize in retrieving or analyzing lots of information, but usually without requiring much judgment or interpretation."
However, those in tech thinking they need to change careers given AI's benefits may be put at ease knowing that the CEO of GitHub, where people can share code, believes ChatGPT can be helpful.
"I think the next generation of developers will be used to AI and it's going to be incredible," Thomas Dohmke, CEO of GitHub, told Computer Weekly. "Technologies such as ChatGPT will enable a new way of learning, so young developers can interact with AI and learn at their own pace, whether it's through tutorials or scripts in a predefined storyline."
Muro noted some finance jobs that deal with manipulating massive data may be at risk. For instance, Muro said that AI can use data from a "financial company to forecast a better investment mix."
"That means that a lot of well-paid financial workers could, in fact, be turned into chatbots," Muro told Insider.
But those worried about their finance job may want to note, as Insider's Paige Hagy and Bianca Chan reported, ChatGPT is "black-boxed," meaning that the underlying processes that lead it to potentially generate investing advice are often not easily interpreted by humans, per Charles Hearn, the chief technology officer at Alloy. Without knowing how the chatbot came up with answers, Hagy and Chan report that it could be hard for this to pass some regulatory requirements.
Some legal jobs could be put at risk as chatbots like ChatGPT become more proficient with legal language.
"The smarter technologies are getting in terms of analyzing, interpreting, and creating language based text – that's actually at the core of what a profession like that is doing," Anu Madgavkar, a partner at the McKinsey Global Institute, said.
AI was already putting some legal jobs on the trajectory to be at risk before ChatGPT's launch. Paralegals, contract lawyers, and court consultants are three of these kinds of jobs noted in a 2019 post from Bloomberg Law about which jobs could be at risk in the future.
Madgavkar told Insider that various kinds of media jobs could be affected by ChatGPT and other AI tools.
But there can be issues with AI doing media work.
CNET tried an internal AI engine for some articles where stories were edited before going live. However, there were still mistakes where "a small number requiring substantial correction and several stories with minor issues such as incomplete company names, transposed numbers or language that our senior editors viewed as vague," according to a post from Connie Guglielmo, editor in chief of CNET.
Another issue in creating stories with AI could be misinformation, as Insider's Samantha Delouya noted. Delouya experimented with ChatGPT, asking it to write a story that Delouya had already reported on about a Jeep factory. But in this experiment, the chatbot made up quotes for Carlos Tavares, CEO of the automaker Stellantis.
"These models are trying to come up with text that is plausible according to their model. Something that seems like the kind of thing that would be written," Conitzer said. "They're not necessarily trying to be truthful."
OpenAI's chief technology officer Mira Murati similarly noted this.
"ChatGPT is essentially a large conversational model—a big neural net that's been trained to predict the next word—and the challenges with it are similar challenges we see with the base large language models: it may make up facts," Murati told TIME.