Until now, the internet has provided a two-way messaging system: users query search engines to find web pages, products and services. Producers send paid ads to reach users. Data trails left by consumers power online targeted advertising that generates revenue that supports many online business models. Google Search, the dominant general search engine, epitomises this strong link between the two messaging channels.
Artificial intelligence services are altering this web navigation system. AI answers to complex queries are derived from original content, doing the work of consulting webpages and saving users time and effort. But this is disadvantageous for online service providers because it re-directs user attention away from original content and from revenue-generating advertising.
AI ‘answer engines’ also complement general search engines. AI is better at answering complex questions that search cannot answer. Search engine providers have started switching between search and answer on the same page, depending on the query. Search engines thrive on data-driven network effects that result in winner-takes-all monopolistic markets, such as Google Search. AI models do not exhibit network effects, though harvesting user data and leveraging search engine data in AI opens that possibility.
AI-induced competition in search supports the efforts of policymakers to weaken Google Search’s monopolistic position, notably by obliging Google to share data with competitors. These remedies have so far not yielded tangible results. Competition policymakers may need to rethink their approaches to Google’s dominant position in search and advertising markets in the face of AI pressure.