David Carmona, from Cádiz, heads the artificial intelligence and innovation division of Microsoft, the company that has most quickly begun to make generative AI available to companies and the general public through its products.
With AI there is a cascade of news almost daily. Will we have a more stable outlook?
In my team we joke and say “what a year of artificial intelligence we have had today”. Even we have a hard time keeping up with all the innovation and research. We are just at the turning point of one of the great revolutions of humanity, at the same level as others such as the personal computer, the internet or the mobile device.
How will Microsoft integrate AI into everyday use?
Our approach is to democratize its use. In any task, with a concept that we have developed, which is what we call co-pilot. It will improve our cognitive abilities to be able to reach a higher level of productivity that we do not have now and new ones to scenarios that we could not cover before.
One of the big problems with AI is that the regulation is still in the drafting phase. What is Microsoft’s position?
We are one hundred percent aligned with the objectives of the European Union. This regulation does not have to be a brake if it is well implemented, quite the contrary. We want to give guidelines of where we see AI going, so that regulation can think not only about today’s technology, but also about tomorrow’s technology and we can have regulation and technology aligned as much as possible. For that you have to go a little ahead.
How do you do that?
We have been working on this new generation of generative artificial intelligence for several years. On the other hand, in 2016 we established our principles of responsible artificial intelligence development. We have acquired knowledge and learning about what has worked, what has not, what can be done better, and what we see working in companies to guarantee responsible development.
That sounds like some kind of Moore’s law – which predicts the increase in chip capacity…
The equivalent of that metric would be the size and complexity of the models. When we talk about the cost of these models and how they scale, you always have to plan for the future.
When you associate AI with quantum computing, it sounds crazy. How far will it go?
With AI you can reason about knowledge from the most basic, which is human language, but there is an even more important revolution to come, which is that they can reason about the language of nature, which is chemistry. By being able to understand nature we will be able to create new scenarios that until now were impossible. Imagine, for example, simulating properties of molecules to create new batteries, materials, drugs to treat diseases, new ways to develop fertilizers for global food production. Two weeks ago we announced the first Microsoft product in this domain, Azure Quantum Elements, which unites two great AI capabilities, that of reasoning about chemistry and, on the other hand, having a co-pilot so that the scientist can reason about all previous knowledge.
How is AI going to help us fight climate change?
The most basic way to deal with climate change is with a revolution in scientific discovery in areas like the development of new energy, batteries, more sustainable materials, and solutions to that problem. It won’t be something overnight.
How far away are we from an AI that has consciousness?
We do not work in that direction. We are working on a long-term artificial intelligence that complements the human, the co-pilot concept.
What will be the rate of adoption of AI in the coming years by the public?
I think what we’re going to see in the next few months, I wouldn’t dare say years, is a more active use of artificial intelligence by the user. From our developer copilot, GitHub Copilot, we have some very interesting data on how it has impacted work. On average, there has been an increase in productivity of 55%. Imagine that you suddenly manage to do your job in half the time. This will reach all professions. There will be no professions that do not benefit from AI and, from Microsoft’s point of view, always increasing our capabilities and not completely replacing the human.
