AI to the Rescue? Breaking Google’s Grip on Search
There was a time when searching the internet meant precisely that. Enter a few words, refine the search, and Google would produce an impressive selection of relevant results. It was so useful – an extraordinary advance, and deserved its runaway success.
Those days are long gone. Search today for a specific product, particularly something relatively obscure, and get ready for a snowstorm of irrelevance. The results are astonishingly bad – just page after page of sponsored products, suppliers and superficially related alternatives. In a bizarre reversal, it now requires more effort to sort through the spam than an old-fashioned catalogue. Google is extraordinarily good at identifying commercial opportunities arising from a search, and simply awful at satisfying the search itself.
Such is monopoly. Google controls almost the entire process: indexing information, ranking results, placing advertising and presenting the answers. From its standpoint, a customer searching for a specific answer and an advertiser paying for generic attention overlap only superficially. A clear choice was made to pivot from delivering few relevant results to selling broad customer preferences to the highest bidder. Without competition, there has been no pressure to reverse that trend.
AI offers an interesting prospect: the return of genuine competition in search, with customer satisfaction rather than advertising revenue once again determining success. While AI's current customer-friendly phase should not be mistaken for permanent generosity, there is a fundamental difference in how its market can operate that offers some optimism.
Consider the oil majors. An industry dominated by a relatively small number of enormous companies supplies an essential input to countless other businesses. Those businesses compete vigorously in how they use that input, without needing to develop their own oilfields or refineries.
AI will likely develop along similar lines – a relatively small number of companies supplying the underlying intelligence, but thousands of independent businesses able to use that capability to develop competing services. An effective search business may no longer need to build Google's enormous infrastructure from scratch. It could combine AI with existing sources of information and compete on the quality, reliability and relevance of the answers. For customers who lost their primacy, the good news is that commercial success could once again depend on satisfying the person searching, rather than maximising the advertising opportunities created by the search.
None of this is guaranteed. AI companies could exploit their dominance, favour their own services and construct new barriers to competition. The oil majors themselves are hardly strangers to vertical integration. However, the important possibility is structural, not behavioural. AI may make it easier to separate the ownership of powerful technology from control over how that technology is used. Even an oligopoly of AI suppliers could support a considerably more competitive market for the services built upon it. Such a scenario would represent a significant change from the integrated business model that has allowed Google to dominate search for so long.
Much of the debate about AI understandably concerns the dangers of concentrating extraordinary technological power in very few hands. Yet the same technology may also undermine some of the established concentrations of power in the digital economy. Perhaps the most encouraging consequence of AI will be not another spectacular technological breakthrough, but the restoration of something rather ordinary: a search engine whose principal objective is to find what was actually being searched for.
This time around, regulators should encourage that competition rather than allow another customer-demoting oligopoly to develop.
