How search engines change with “intelligent” chats: new rules of the game

This is the slightly expanded version of the comment published in Il Sole 24 Ore on February 8, 2023, after the announcements by Google and Microsoft on the convergence of automatic chats and search engines.

How to talk about books you haven’t read? For the cultured person it is possible and legitimate, according to Pierre Bayard, psychoanalyst, author of a book entitled that question, a true bestseller from around fifteen years ago. So what’s wrong with ChatGPT writing about things it didn’t understand and know nothing about? OpenAI’s generative AI that could write for humans has, for now, only prompted many humans to write about AI. In reality, on a cognitive level, that artificial intelligence is rather in its infancy: those who ask it to talk about issues it does not know are usually surprised by its eloquence, but those who ask it about topics they know discover that they do not have the very clear ideas and fills his answers with errors. Well: could the announcements of these days change the scenario?

Indeed, Alphabet and Microsoft seem to want to connect their search engines, Google and Bing, to their generative artificial intelligences, respectively Bard and what until yesterday was called ChatGPT. With what consequences?

At best, artificial intelligences should be able to better inform themselves by exploiting the organization of knowledge developed by search engines and thus becoming more competent. Furthermore, there will be a redefinition of the interface with which search engines are queried, which will become a conversation with questions in natural language and answers emerging from the aggregation of different sources of knowledge, published in separate boxes from the traditional list of recommended sites .

To obtain these results, one must definitely innovate with respect to ChatGPT. The chat that has gained so much attention does not base its answers on particularly up-to-date information: it cannot use search engine answers, for example, but works on a monolithic corpus of data, which stops at 2021. Bard, Google assures, instead, it is already capable of using the results of the search engine, even if the product is still experimental. Which evens the score with the new Bing which should have a conversational interface with updated answers, thanks to a new artificial intelligence, GPT 3.5. All of this will probably have spectacular effects. One can imagine that questions to the search engine can be refined by dialoguing with the chat and then obtaining answers composed of elaborations of information contained in a series of sites retrieved by the engine. Interesting, but not without problems. One thing is that the engine returns a list of sites and the user chooses the ones that suit him. It is another matter that it returns a report instead: it will not be easy, in that case, to understand if the results produced by artificial intelligences contain errors, prejudices or misinformation.

You can bet that the search engine giants, especially Google, will tread lightly. Although maybe they will create tools capable of attracting attention. It is even naive to point out that all this enthusiasm for generative artificial intelligence is partly linked to promotional strategies of the giants involved. It is noted that Microsoft appears to be on the attack while Google defends its core business. But there will be consequences for the ecosystem: there will also be investments, start-up funding, stock market placements, some speculation and many efforts by users and customers to distinguish wheat from chaff. It takes getting used to looking beyond the glitter of automatic replies to understand how they are formed and how reliable they are.

Meanwhile, regulation will need to speed up to prevent potential harm. Answers will have to be found on the copyright of the original texts from which artificial intelligences derive their results, on the responsibilities for possible plagiarism and errors, on the forms of automatic comparison between information sources that may contain prejudices, and so on. The European Commission is moving forward with its AI Act. And it has a lot of work to do.

Photo: “Artificial Intelligence – Resembling Human Brain” by deepakiqlect is licensed under CC BY-SA 2.0.