An artificial intelligence that knows how to reason? The idea of Google (pianetacellulare.it)
Agentive capabilities is a phrase we often hear in AI conversations, but do you know what it means exactly?
Every time there is a small, big technological revolution, the way we talk about technology also changes. Between terms that are emerging as new technical jargon is the expression agentive capabilities.
This is an expression linked precisely to artificial intelligence. Something that should tell us what is the ultimate purpose of these new tools. In particular to understand what Google wants to do with its Gemini project. The expression agentive capabilities was in fact used to explain how the new Google Search works.
Why talk about agentive capabilities with artificial intelligence
The idea of artificial intelligence is already an elementary and at the same time very complex concept. In fact, while neurologists and experts in the study of the human mind have not yet reached the crux of the matter (in fact we do not know how part of our brain works), other scientists seem to have managed to replicate a part of the functionality inside a machine.
For this reason we talk about intelligence artificial. That is, something non-organic that is able to behave, in some contexts, like human intelligence. In order to arrive at the results that we see in ChatGPT and the others, the artificial intelligences were subjected to very long tests in which they learned.
However, Google’s purpose is slightly different. Instead of just giving to users a parrot repeating what he has learned by heart, Gemini will be characterized precisely by agentive capabilities. The literal translation is agentive capabilities.
Something that doesn’t exist in Italian. Wanting to give a clearer translation, agentive capabilities means ability to act (think) autonomously and proactively. We are still far from the silicon brains that think and dream of electric sheep but we are certainly a step in that direction.
What, for now, is the way in which agentive capabilities is expressed in Google Search with Gemini is the possibility of handle long and complex questions without having to ask question after question. In the Gemini demo presented during the Google I/O 2024 event, for example, a 7-day meal plan with recipes to be made in the microwave was asked for. The AI responded with all the information in a single search.
Are we ready to rely on AI?
Ask Google something complex and get a sensible answer. Solve doubts in seconds. See yourself organizing an entire holiday based on your requests. Change your lifestyle and try to cook vegan for a week without having to browse through forums and portals. All of this, Google tells us, is and will be possible using Gemini and its agent capabilities.
The real question then becomes whether we are as human beings ready to leave the command to artificial intelligence a bit. Above all, to trust the answers we are given. In addition to learning new words we will also have to learn new ways to interface with technology.
