Skip to content

Kirk LaPointe: Google’s AI is delivering a blow to the newsrooms it's built on

As AI summaries in search results devour web traffic, news publishers must move beyond derivative reporting to attract and sustain audiences
google-search-phone-credit-bongkarnthanyakij-istock-gettyimagesplus
As Google's AI Overview hurts newsroom web traffic, publishers must prioritize original journalism and distinctive reporting, argues Kirk LaPointe.

Thanks in advance for taking the time to read this column. Not sure how you found it, because a new threat arising from artificial intelligence’s advances is rapidly making this, and other sources of journalism, easier to avoid.

The latest challenge for news organizations is technology that answers your search for information with an AI-driven summary—what Google calls an AI Overview—atop the screen’s display of the most relevant links to a query.

Studies show many major news outlets are experiencing serious declines in online traffic in recent months as AI Overview takes hold, because the summary somehow satisfies readers sufficiently so as not to click through to the journalism website sources.

The consequences of this development are profound. If you don’t believe me, read what the prominent AI platform ChatGPT—essentially, one of the perpetrators—had to say about AI Overview’s impact when I asked about potential ramifications: “The threat posed by Google’s AI Overview to news organizations is serious and potentially existential, particularly for publishers that rely heavily on search referral traffic for ad revenue, subscriptions, or visibility.”

For many years now, online journalism has earned a large share of its readership—about half, in fact—from individual searches, primarily using Google.

For decades, search engines and social platforms didn’t pay news publishers for helping them attract users and provide journalism to consume—although, to be fair, publishers benefited from the technology that drove traffic to news so they could display advertising.

The technology stack’s capacity to commandeer digital advertising at the expense of conventional news organizations became the driving purpose behind the federal Online News Act. It compelled Google—and tried to compel Meta—to compensate publishers for a share of what they earned when users of their platforms linked to news websites. Google reached a deal to spend $100 million a year in Canada doing so. Meta walked away from any negotiation, and scrubbed the links to news sites from Facebook and Instagram rather than participate.

But it took all of about a year for the technology to render the legislation out of step. The disruption was far from over, adding to issues involving the broad challenges for the sustainability of news.

As generative AI became accessible to all, several publishers took legal action against AI firms for using copyrighted news as a source of data without paying its creators. Most could see, though, that it was only a matter of time before generative AI would enter a field previously occupied by journalism.

As Google put it when I searched for its explanation of AI Overview: “These overviews utilize generative AI to synthesize information from various web sources into concise summaries, offering users a streamlined way to grasp key information and explore related content.”

It states that AI Overview also is “making it easier for users to understand complex topics and discover diverse websites.” Sure, the first of those two points is true, but the second of the two seems not—at least yet.

This week, two prominent founder-publishers of independent Canadian media outlets wrote jarringly about the effects of AI Overview on the industry’s ability to sustainably finance news. David Skok of The Logic and Rudyard Griffiths of The Hub (full disclosure: I am one of its contributors) sounded an alarm about how the already-challenged environment for journalism stood to suffer if left unaddressed.

“What’s urgently needed is a fundamental rethinking of how journalism is valued, accessed and supported,” Skok wrote. “This will require co-ordination between publishers, platforms, policymakers and, yes, the public.”

Griffiths’ organization is participating in tests of the pre-market version of Google’s AI Mode that “merges ChatGPT’s prompt functionality with Google’s near real-time, all-encompassing search.” He wrote: “From what we have experienced … it should terrify news execs.”

Now, neither publisher is waving the white flag.

“Rather than ‘killing’ journalism, AI-powered news distribution could end up saving news by prompting a much-needed burst of innovation and genuine competition,” Griffiths said.

Like Griffiths, Skok suggests that his outlet could be one that “ultimately benefits from this change because our original journalism stands apart from the commoditized information that chatbots mostly tend to regurgitate.”

What they and others suggest is that newsrooms will need to cover less and uncover more—there is far too much undifferentiated, duplicative, pack-mentality stenography in the craft—and that the rewards ahead will go to those taking the unique paths to discovery in telling us what we didn’t know.

“Necessity, as the saying goes, is the mother of invention,” Griffiths said.

Clearly, though, if credible source material for AI can’t be sustained financially, the quality of the AI summaries will diminish, so it will be interesting to see how governments globally contend—and if they intervene—with the new dynamics of information consumption.

Meantime, out of bald self-interest, let me say: bookmark your regular sites, and next time you search for information, scroll down and click the sources of the result. There is more to the iceberg than its tip.

Kirk LaPointe is a Lodestar Media columnist with an extensive background in journalism. He is vice-president in the office of the chair of Fulmer & Company.