Since ChatGPT was launched in November 2022 a new form of information retrieval has been available. Before the release of Large Language Models, search - a market dominated by Google - was the fastest way to find information online.
I still use Google numerous every day, and haven't previously been tempted to use Perplexity much, which has found a keen userbase amongst AI enthusiasts. I found it's use cases too niche to change my behaviour, and one LLM subscription plus Google still felt like more than enough.
ChatGPT search arrived on 31 October. In many ways, this is a big moment in how the search market could be shaken up by new Artificial Intelligence products. I tested it out over the weekend and am impressed by how quickly it can return information and provide citations.
Earlier iterations of ChatGPT could not pull real time data from the web, so for many this will be a welcome update.
Hype and challenges
Of course, almost immediately I saw numerous posts on LinkedIn about how Google was now dead. I have written an longish post (prior to ChatGPT search’s launch) on why I feel the gulf is still too vast to see this happening. In summary:
While ChatGPT is the market leader in chatbots, at 80% market share, the LLM market is not big enough yet to seriously challenge Google. There are roughly 300 million regular users of LLMs, while search is pretty much ubiquitous to everyone with an Internet connection.
Of the search market, Google has just over 90% market share - it is a monopoly. Read Johnathan Taplin's Move Fast and Break Things for more on that.
Alphabet (Google's parent company) makes vast profits - probably north of $75 billion in 2024. ChatGPT, while backed by other big tech companies, makes a $5 billion loss.
Added, killing Google will take a huge shift in default Internet user behaviour. Such ingrained practise will not disappear overnight.
Google problems
All of that said, there will still be alarm bells going off at Google about how to face the new AI based search threat. Google issued a 'Code Red' on AI back in February 2023 around how it could better meet the fast moving developments.
My understanding of search shifts since then is that Google has created a paradox on its main product. While it needs to ensure its index maintains its quality, against a potential tide of AI generated content, it has also given an immense preference to Reddit - which is user generated.
Further, its AI overviews have many in a spin. Many are claiming (without much current evidence) that it will lead to a reduction in traffic to websites. Some are vague or inaccurate, while there are straightforward copyright concerns. It appears that it's in a half in-half out situation of having a huge legacy product that needs to feature AI in some way.
Copyright and AI summaries
I think generative AI will soon be faced with a copyright problem. The search citations require chatbots and AI overviews to provide summaries, most of which are unlicensed. Microsoft CEO Satya Nadella has called for a rethink of copyright laws to redefine fair use. Of course, Microsoft has bought a stake in Open AI, but I think this is a very challenging position.
Media publishers have long relied on Google for traffic. It is my view that is ChatGPT had much more search market share, referrals from search would go down. A generic search for the US election provided me with citated information from three sources, and one summary is from the publisher I am currently working with:
Early voting has commenced in several states, with a significant number of ballots already cast. However, due to varying state regulations on processing early votes, the final results may not be immediately available on election night. For instance, Pennsylvania's vote-counting process, especially concerning pre-poll votes, may take up to two weeks, potentially delaying the overall result announcement. The Telegraph.
I don't feel there's much incentive for a ChatGPT user to click through to the citation. This issue will potentially cause a big problem to publishers, and without maintaining a vast amount of licensing deals, lawsuits will be incoming. The unintended consequences is another potential round big tech dystopia, where the winner of the AI search wars takes all.
Writing in the MIT Technology Review, Ben Brooks makes a similar point:
At its best, AI search can better infer a user’s intent, amplify quality content, and synthesize information from diverse sources. But if AI search becomes our primary portal to the web, it threatens to disrupt an already precarious digital economy. Today, the production of content online depends on a fragile set of incentives tied to virtual foot traffic: ads, subscriptions, donations, sales, or brand exposure. By shielding the web behind an all-knowing chatbot, AI search could deprive creators of the visits and “eyeballs” they need to survive.
So what is going to happen in the AI search wars over the long term? It's difficult to say with much certainty, but I don't think Google's stranglehold on the search market will be threatened any time soon. The bigger issue is if Google moves in this direction. It is most likely to be cautious on this, due to the threat of lawsuits and regulation.
Potentially the game will change significantly with the introduction of autonomous agents - something that Claude announced on 22 October. This, to me, seems the bigger potential shakeup. Eric Schmidt, in a now deleted Q and A at Stanford, said their arrival will be 'much bigger than the horrific impact we’ve had by social media.' I'm not sure social media's impact can be defined as 'horrific' - but it has been disruptive in many ways outside of marketing and business. Autonomous agents, meanwhile, may challenge the dominance of Google in entirely new ways. How quickly they arrive, and under what regulation, remains to be seen.
