Former Google news boss predicts chatty AI models like ‘ChatGPT’ will replace journalists

According to Jim Albrecht, large language models pose an existential threat to journalists and the global news industry.

What if, every time you read an interesting news article, you were given immediate access to the journalist who wrote it so you could ask follow-up questions and discuss the material?

That’s what former senior director of Google News Jim Albrecht laid out in a recent op-ed for The Washington Post.

Artificial intelligence journalists

Per the article, Albrecht believes modern large language models (LLMs), such as ChatGPT, will radically change the news industry, much like the advent of the world wide web did so in the early 2000s.

Before the internet, the average person got their news from television and print news media such as newspapers and magazines. While these formats still exist today, their function has been diminished by the emergence of web-based news and information sites.

Gone are the days when the quickest way to check a sports score or find someone selling an armoire in your area was to flip to the appropriate section of your local newspaper. Now, it’s a simple matter of punching in the correct website and clicking on a link.

However, as Albrecht writes, the web may have changed everything about the way news articles and similar content are distributed, but “the articles remained mostly untouched.”

A new paradigm for news

Untouched until recently, that is. In November 2022, OpenAI launched its ChatGPT LLM service. At first, users could prompt the bot to drum up information from nearly any source prior to 2021 (a technological limit due to the cutoff date of the pre-training data used to build the model).

Fast forward just a little over a year, and ChatGPT (and similar models) can access news articles currently available on the internet, even if they’re hot off the virtual presses. More importantly, LLMs can reword and rephrase information from existing sources and deliver aggregate news to users in a single interaction space.

Albrecht believes that this paradigm shift will lead to another revolution in journalism. This time, rather than moving the media to a more accessible platform, news readers will shift their attention from static news articles to interactive journalism experiences led by ChatGPT-style chatbots.

Per the op-ed:

“The users will interact less and less with the actual articles and instead talk about the articles with what the tech industry used to call ‘intelligent agents.’”

Albrecht predicts that news readers of the future could prefer to get their news from AI sources capable of “talking” to them about the articles over humans, who are physically incapable of interacting with millions of people simultaneously in what’s ostensibly a “one-on-one” conversation.

The end of modern journalism?

It bears mentioning that Big Tech has predicted the end of traditional print/online journalism since 2015, when Facebook pivoted to video. By 2017, however, many of the newsrooms that attempted to force the transition had found that reports of the death of online print media had been greatly exaggerated.

In the time since, ChatGPT and its ilk have risen to prominence, and innumerable news outlets have run experiments using AI to generate articles with mixed results.

While some of these trials have produced successful results — especially those operating with full disclosure — many have resulted in embarrassment for the organizations using them as mistakes, and a lack of disclosure alienated established readers.

Related: Legal experts weigh in on landmark NYT vs. OpenAI and Microsoft lawsuit