A different version of this article has been published in Italian on Crossroads, my blog at Il Sole 24 Ore. The Forum Network is a space for experts and thought leaders—from around the world and all parts of society— to discuss and develop solutions now and for the future. It aims to foster the fruitful exchange of expertise and perspectives across fields, and opinions expressed do not necessarily represent the views of the OECD.
How to talk about books you have not read? For the educated person, it is possible and legitimate, according to Pierre Bayard, a psychoanalyst and author of a book with that question as its title, a real bestseller some 15 years ago. So what's the harm if ChatGPT writes about things it doesn't understand and knows nothing about? OpenAI's generative artificial intelligence that could write instead of humans, for now, has only prompted many humans to write about artificial intelligence. Actually, on a cognitive level, that artificial intelligence is rather novice: those who solicit it to talk about matters they do not know are usually surprised by its eloquence, but those who question it on matters they know well find that it is not very clear in its ideas and fills its answers with errors. Well: could the announcements of these days change the scenario?
There are direct consequences affecting users of the search engines. And there are indirect consequences affecting the whole digital ecosystem.
Indeed, based on the announcements of the past two days, 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? There are direct consequences affecting users of the search engines. And there are indirect consequences affecting the whole digital ecosystem.
Also on the OECD Forum Network: The Green and the Blue: How AI may be a force for good, by Luciano Floridi & Anna Christina (Kia) Nobre, University of Oxford
As for the direct consequences: at best, the artificial intelligences should be able to better inform themselves by exploiting the knowledge organisation developed by the search engines and thus become more knowledgeable. This could lead to better chat sessions. In addition, there could be a redefinition of the interface for queries on search engines: it could become a conversation with natural language questions, while answers could emerge from the aggregation of different sources of knowledge, published in boxes separate from the traditional list of recommended sites.
To achieve these results one must definitely innovate from 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 body of data that stops at 2021. Bard, Google assures us, on the other hand, is already capable of using search engine results, although the product is still experimental. That evens things out with the new Bing, which is expected to have a conversational interface with updated answers, thanks to a new artificial intelligence, GPT 3.5. All of this is likely to have spectacular effects. One can imagine that one could refine questions to the search engine by conversing with the chat and then get answers composed of processing information contained in a series of sites retrieved by the engine. Which is interesting, but not without problems. It is one thing for the engine to return a list of sites and for the user to choose the ones that are right for them. It is another account if it returns a compiled text instead: it will not be easy, in that case, to understand whether the results produced by the artificial intelligences will contain errors, biases or misinformation.
You can bet that the search engine giants, particularly Google, will be careful, although they will create tools capable of attracting attention. That all this enthusiasm for generative AIs is partly related to promotional strategies of the giants involved is even naive to point out. It is noticeable that Microsoft appears to be on the attack while Google defends its core business.
We need to get used to looking beyond the glitter of automated responses to understand how they are formed and how reliable they are.
But for the digital ecosystem there will be no shortage of consequences: there will also be investments, funding in start-ups, stock market placements, some speculation and many efforts by users and customers to distinguish the wheat from the chaff. We need to get used to looking beyond the glitter of automated responses to understand how they are formed and how reliable they are. The whole “search engine optimisation” (SEO) business will have to adapt. Advertising will mix with regular information in different ways. Trustability of search engines and their service will change. The business of comparing different commercial offers will change profoundly. Competition between giant and small innovative companies will also be different.
Meanwhile, regulation will have to accelerate to prevent potential harm. Answers will have to be found about the copyright of the original texts from which the artificial intelligences derive their results, liability for possible plagiarism and errors, forms of automatic comparison between information sources that may contain bias, and so on. The European Commission is moving forward with its AI Act. And it has a lot of work to do.
To learn more, find out also about the OECD's work on Artificial Intelligence