Earlier this week, we learned about the Department of Justice’s plans to end Google’s monopoly over internet search, and more specifically it’s plan to try to restrain Google from using that monopoly to influence the development of artificial intelligence.

That got us thinking. A week from Saturday will mark two years since the debut of OpenAI’s ChatGPT. So how exactly has AI developed since then? And after two years of all the hype and bans and scary open letters warning of the end of humanity, how much are people and businesses actually using this stuff? 

I hesitate to tell you this, but Google has a relatively new AI product called NotebookLM that can basically turn anything you feed it into a podcast.

Economist Anton Korinek at the University of Virginia uses it all the time, mostly to summarize research papers that are a slog to actually read.

“I sometimes do that now because honestly those voices they’re engaging, it’s so well done, that’s its quite impressive,” he said.

Notebook LM is just one of the many leaps the tech has made over the past two years. Generative AI is also less likely to just make things up.

And yet, AI is not being adopted as quickly as Korinek would have thought.

“There is this almost free tool that’s extremely powerful and its like $100 bills lying on the floor and nobody’s picking them up,” he said.

Convincing business to pick up those low-hanging dollars is part of Shane Cumming’s job. He’s the president of Acrolinx, a company that offers AI content standard services to big enterprises.

Cummings said while tech and finance have been early adopters, other industries are just more risk averse.

“One of the one of the things that hasn’t changed in the last two years is that large enterprises, all enterprises, they are accountable for the content they produce,” he said. “They’re legally accountable, they’re ethically accountable, they’re financially accountable.”

Yet about 40% of the U.S. working-age population use generative AI in some way, according to a recent study co-authored by Vanderbilt University economist Adam Blandin.

He said the early spread of AI mirrors the pace of a previous tech revolution.

“Our overall adoption rate among workers looks very similar to the overall adoption rate of computers in 1984. Just to give you some perspective 1981 was when the IBM PC was released,” he said.

Blandin also said there may be more AI use going on in the workplace than we know about. It may be a few more years until employees are comfortable telling their boss they used ChatGPT for that Powerpoint.