#5: AI Search Chatbots And The Publishing Industry
Microsoft and Google are battling for the search market and how we interact with the web. What will happen to the media if search engines ditch the results page with multiple links and switch to an endless conversation with a chatbot?
1. AI won't kill the news media, and its reporting will get regulated. AI isn't good with facts. The media knows it, and the public notices it. Progress will be made, but trust in AI's accuracy will erode. Come next presidential election in the US, the legacy media and the EU regulators will try to limit what Google's or Microsoft's bots can tell you about candidates, public health, and other sensitive topics.
2. AI won't reduce political polarization. Brands influence our perception of information. You are more likely to dismiss a piece by a media outlet the further it is from you on the political spectrum. A search chatbot will not show you the publication’s logo the second you ask a question. In theory, this will make readers more receptive.
It is, however, improbable that the AI will be considered less biased than NYT or Fox. The left and the right will want to know who and how instructs Kenyan outsourcers labeling the data. The answers won't satisfy anyone, and chatbots will be pronounced Republican or Democrat.
3. AI will drive the valuations of media businesses. Pressured by their shareholders, all public media companies will have to have an AI strategy. Media startups will try to ride the wave of AI funding. Almost all in-house AI products (including proprietary AI chatbots) won't impact the core business or provide a competitive advantage. However, it will take years for the market to understand that AI-infused BuzzFeed is still in the content business, no matter how many AI quizzes Buzzy The Robot will generate.
Ten years ago, many media companies were trying to rebrand themselves as tech companies by selling their proprietary content management systems (CMS) on the market. Only The Washington Post succeeded (Arc XP drives about $40-$50 million in ARR), showing that for media, tech is hard and capital-intensive.
4. AI will hurt the media's e-commerce revenue. Journalists sometimes frown upon comparing toaster models, but e-commerce content (also known as affiliate and shoppable) is honest work on the $17 billion market. It generates reliable commission revenue on every sale of the compared products and depends almost exclusively on search traffic of users with high intent (those who type "the best toaster" in the search bar).
What it lacks is a moat against Big Tech. The legacy media has a defense in news reporting due to higher trust and regulator pressure. But for Big Tech, the downside of recommending a suboptimal toaster is incomparable to the downside of getting the results of an election wrong. Big Tech will be comfortable with "moving fast and breaking things" on this one, starting with providing product recommendations without any links to the legacy media e-commerce articles. You can't have too big of a share of the e-commerce market, even if it means selling a lot of bad toasters.
This lack of defense will make droves of mid-sized Wirecutter-like publications to sell or close, giving them a very narrow window to react, considering that most get 70%+ of visits from search engines.
5. AI will help the big players get even bigger through partnerships with Big Tech. Big Tech will use such partnerships to shield itself from regulators. In this case, it’s not Bing that will give you meaningful info on vaccines, but Nature, as a source provider for Bing's chatbot.
Big Tech will try to solve the trust issue, and the media business will try to leverage these relationships to protect the revenue streams most affected by the rise of conversational search. This trend will accelerate when more players enter the scene and try to carve a more high-end niche (think Apple's search engine).
We look forward to this new AI-driven transformation of the constantly doomed publishing industry. And if your media outlet is just about reviewing toaster models for e-commerce revenue, the latest time to switch your business model to subscription was probably a couple of years ago.