Cold Comfort for Publishers: Consumers Slow to Embrace AI Search

Artificial intelligence may be evolving fast, but U.S. consumer attitudes concerning the technology are moving much slower.

That’s evident in new survey data provided exclusively to VIP+ by decision intelligence company Morning Consult that indicates little has changed with regard to generative AI-powered search engines in the period between February and May — the months in which usage of and interest in AI has exploded and search giants Google and Microsoft have begun deploying chatbot functionality on a limited basis.

Audience inertia could offer some solace to online publishers, many of whom fear conversational search could choke off the flow of traffic to their websites, triggering an extinction-level event for their industry.

News Corp. CEO Robert Thomson warned Google and Microsoft chatbots could “fatally undermine journalism” in recent comments at the International News Media Association’s World Congress. At the Semafor Media Summit in April, IAC chairman and senior executive Barry Diller said that “there will be no publishing” if conversational AI proceeds with no limits or compensation.

Since it seems unlikely that helpful regulation will form in the near term, the industry will likely be on its own to defend against and negotiate with AI companies. Determined not to repeat the mistake of failing to anticipate the disruption tech giants have delivered to publishing in decades past, execs are moving proactively and assembling task forces to assess the right response to both the challenges and opportunities that await them.

Earlier this week, the trade organization Digital Content Next (formerly Online Publishers Association) released its principles for the development and governance of generative AI, among them protecting proprietary content under existing copyright law and seeking fair compensation for the use of their IP by web-scraping AI systems.

In April, fellow publisher trade organization News/Media Alliance put out similar principles. Some publishers — including Bustle Digital Group and Leaf Group — are reportedly even heading off the threat by reducing reliance on SEO-driven content.

Meanwhile, Microsoft and Google are proceeding to roll out the new tools more publicly. In February, Microsoft launched Bing Chat, its AI chatbot powered by OpenAI’s GPT-4 large language model (LLM), in a private preview for waitlisted users. As of May, Microsoft made Bing Chat available for public sign-up.

In March, Google launched Bard in the U.S. and U.K. with a waitlist and in May opened it up to over 180 countries and territories. Also in May, OpenAI’s ChatGPT integrated Bing to support its own search functionality, initially available to ChatGPT Plus subscription users and soon rolling out to free users through a plug-in.

Bing Chat, Bard and ChatGPT’s search option each show citations as embedded source links in the text responses they generate. But these citations in AI outputs aren’t focal parts of the results, so it would seem less likely users will actually click into publisher links these systems provide.

In effect, chatbot search engines increase the risk of “zero-click” search, where the user receives information without needing to click into a publisher site. But the ability of ad-supported publisher content to reach monetizable audiences at least partly relies on search. Publisher traffic isn’t solely search driven, but search remains among the dominant ways internet users locate and access information online.

But for now at least, users aren’t flocking to AI search, which isn’t taking over for typical online search behavior or intent, and perceptions haven’t changed since February. Overall, most (52%) of U.S. adults still don’t intend to switch to a generative AI-powered search engine, versus 54% in February, according to Morning Consult data from February and May 2023 provided exclusively to VIP+.

Consumers are also still more likely to prefer traditional search over generative AI-powered search for accessing information, including believing typical search is best for getting in-depth information about current events (44% of U.S. adults) and providing the most accurate and factual information (42%), per Morning Consult. That’s compared with 30% and 33%, respectively, who said AI search would be best.

Compared with February results, those preferences haven’t substantially changed.

It’s likely that AI search has a trust problem, especially when it comes to sourcing reliable information: 40% of U.S. adults wouldn’t trust an AI search engine to avoid gathering content from websites that promote misinformation, versus 25% who would.

That’s where publishers could find an opportunity and rise to meet a user need that seems only more likely to sharpen. The accuracy of publisher content could be how it best differentiates from AI, as well as from AI-generated synthetic media expected to flood into the online information ecosystem in what the New York Times publisher A.G. Sulzberger recently referred to as a “torrent of crap.” But as AI technology evolves like Godzilla, it’s unclear how long accuracy will be its vulnerability.

Meanwhile, the industry alarm is understandable, and the business model risks are clear. But AI has taken on two faces for publishers — both friend and mortal enemy. Even as chatbots emerge, several publishers, including CNET, BuzzFeed, Insider and the Arena Group, are using AI to help them create content for audiences or create chatbots on their own websites.

Of course, it's still early days for search engine chatbots, with each major one being released to the wider public in May, so it’s likely still too early to detect a real shift in attitudes or behavior. Even so, the latest data should provide some comfort to the publishing industry. With search, which hasn’t substantially changed in its basic functionality since the early days of the web, user habits and mindsets will probably take time to catch up to AI.