Google CEO Sundar Pichai announced this week that his company’s Internet search, the most widely used in the world, is about to undergo a change. Instead of just receiving links to web pages and some suggested answers when making a query, it will now be aided by artificial intelligence. According to Pichai in a Google blog post, “You’ll soon see AI-powered features that provide complex information and different perspectives in simple formats so you can quickly understand what the big picture looks like and learn more about the web.”
The next day, Microsoft CEO Satya Nadella announced that his company’s search engine, Bing, would do the same, using technology from the popular AI chatbot ChatGPT. Microsoft sees artificial intelligence as an opportunity to close in on the leader, as so far no search engine has really challenged Google’s dominance.
However, the new chatbots are not really intelligent. The technology that powers them is called extensive language modeling, which is software capable of extracting related words from a large database and producing complex text and visual art with minimal assistance. Language-model chatbots might provide more human-like responses than regular searches, but they have a problem: they are “dilettantes,” meaning they have “no real understanding of the world” and are “unable to justify their utterances by referencing supporting documents in the database on which they have been trained.”
According to the article, more authority and transparency are required for an AI chatbot to be effective in search. It needs to remove any bias from its training database and teach it to incorporate diverse perspectives. If this is achieved, the bot will move from being an “amateur” to something more akin to an expert.