Google AI is Search
I'm trying not to make a partisan statement when I write that Gemini (Google AI) is also the searching feature. In the news, featured stories about Google restricting search when AI is banned has circulated my feed. New York Times reports that Google now provides schools two options: AI overview or student guidance, a student curated form of Gemini. Advocates who want to use less or no AI have said that having no option without AI is the same as no option and that Google is restricting access to something that has been solved by any other search platform with and without AI.
If plainly the purpose of Google is to find the immediate artifact for its query, then the user relies on Google as a sort of meta-expert, parsing big data to find relevant information. Thinking of Google as a separate product from Gemini goes against the nature of an AI expert system: "expert decisions or recommendations based on an extensive body of knowledge" (AI Ethics, Coeckelbergh). The meta-expert Google must also have the capacity of intelligence: "to do the right thing at the right time" (Oxford Handbook of Ethics of AI, Bryson). This done with conditional statements is further developed by learning language models, which was trained on the big data of the Internet indexed: Reddit, Facebook, websites, etc.
Though an imprecise description of Gemini, the basis of generic AI development originating from a deterministic system based on branching conditional statements to expert systems to machine learning in big data to its chatbot form, leads me to believe that the AI portion of Google is inseparable from its chat.
Search augmented with AI is what we expect from a Google result. Pattern matching, context awareness, and generative language is what has provided us the accurate information result on the query what time does a shop open, how to get home, etc. Without these capabilities, the Google product would be pulling matching strings from a large list and little curation.
Such products exist, but they are designed to avoid the curation that users rely on in Google queries. Using no conditional reasoning on pattern matching and context awareness provides an unfamiliar result, which users may find less useful if they were unaware of the application's limitations. It's my belief that the Google search feature hasn't used this design for many years. Therefore, schools which have banned AI usage in schools, should be prepared to use search tools differently. Less reliance on consumer products may be the answer -- and an active development in tools that use older AI standards (without generative properties and require less energy use).