European creativity in Artificial Intelligence

Image: Ron Lach
Whatsapp
Facebook
Twitter
Instagram
Telegram

By AFONSO FERREIRA*

The European Union should take advantage of the waves created by the democratization of Generative Artificial Intelligence models

What type of research in Artificial Intelligence should be funded in the European Union's next Research & Development (R&D) program, the Frameword Program 10,[I]? When reflecting on this question, I thought it appropriate to ask a generative artificial intelligence system for help. Unfortunately, I had never heard of FP10. So, here is a text that I produced myself.

While I'm not made of silicon, I do have some relevant experience to give an opinion. I coordinated my first project Frameword Program in the early 1990s, at the start of a career that experimented with both sides of the European funding desk – as a researcher at the CNRS in France[ii] and as responsible for programs and policies at the European Commission. Based on this and subsequent work, I believe that my main recommendation for Frameword Program 10 should be to avoid chasing what US tech companies just did. Instead, the European Union should try to finance what they do next. Or after that.

Let me give you an example of how the European Union should not do this. Back in 2000, when Google was launching its world-changing search engine, the Commission was busy funding Frameword Program in circuit-switched networks, thanks to lobbying by incumbent telecommunications operators from some large European Union member states. Later, the European Community woke up to the fact that the digital future would be based instead on packet switching and something called the Internet Protocol. It was then that European Union leaders decided that Google required a European competitor and began calling for projects in – wait for it – search engines.

But by then Big Tech had already moved on. They were building very large digital infrastructures, in the form of huge data centers, to collect all the data they could and then some. IBM, Google, Facebook, Amazon, Baidu, Tik-Tok, and others have begun using this data to advance the scientific state of the art in machine learning and other fields. What about the European Union at this point? Proudly financed projects in digital infrastructure and data processing.

Now the buzz is about Generative Artificial Intelligence (AGI)[iii]. And my conclusion: No. Frameword Program 10 should not put all its AI funds into the late development of a European counterpart. Instead, let's think ahead of the curve. It's time for peace, drugs and rock&roll.

Artificial intelligence for peace

For example, we could start by focusing on peace. The European Union currently has two raucous wars knocking on its door. What about financing projects that aim to use Artificial Intelligence for research and innovation that promote peace between people? After all, there is a lot of data about periods of war and peace in history and their dynamics. There are also many theories of war and peace, as well as studies on the origins of conflicts.

Thus, the Frameword Program 10 it should call for intelligent ways of using Artificial Intelligence to solve optimization problems that try to minimize the probability of war or maximize the probability of peace, based on parameters such as dialogues and other practices to avoid conflicts and new ones yet to be discovered. Of course, this will not be easy, as they are multi/inter/transdisciplinary by default; and finding good performance indicators is certainly a challenge. But I'm sure it will be easy to find many experts willing to contribute to such a different goal. With a nod to John Lennon: all we're saying is give peace a chance.

Drug development is another area that can greatly benefit from machine learning tools. Humanity will benefit even more from new medicines that can be developed more cheaply and quickly. As Google likes to advertise, modeling how proteins are structured is a very difficult problem. Of the approximately 200 million proteins, only 170 thousand were modeled, mainly by very intelligent groups of doctoral students, at the time when the Alpha Fold Of google[iv] was developed. Then the Alpha Fold exhibited such a high level of accuracy in proposing protein structures that the community considered the protein folding problem solved. Now, again according to Google, there are 1,7 million scientists working on the Alpha Fold and more than 18.000 citations of him in scientific literature.

Of course, there are already large drug-related databases that can be used to train new models, but they are mainly owned by large pharmaceutical multinationals. Incentives would be needed for these companies to cooperate with each other and with academic researchers. And there are some prominent and high-reward fields of drug development in which artificial intelligence can change our lives, such as aging research. O Frameword Program 10 there should be an Artificial Intelligence program for the development of medicines, with a strong line of financing for projects related to combating aging and increasing human life expectancy in good health conditions. This has less to do with wanting to live forever than with wanting to be young forever.

Let's let a thousand projects flourish

Finally, the European Union should ride the waves created by the democratization of Generative Artificial Intelligence (IAG) models that has happened since OpenAI revealed its IAG in November 2022 and since Facebook's IAG model, Calls[v], leaked. Today, anyone can develop their own Generative Artificial Intelligence models on laptops. This is the power that begs to be harnessed at the level of a Frameword Program.

With these tools, the 10th Frameword Program should appeal to a myriad of small, goal-based research projects powered by Artificial Intelligence. There should be both calls without specific themes, leaving candidates open to idealizing areas of application, and calls with major, very ambitious and pre-determined themes. In addition to the challenges of peace and the pharmaceutical industry that we just mentioned, these big themes could be in biology.

For example, what makes a collection of liver cells behave like a liver? What is consciousness? There could be major projects on social challenges: how to save humanity from the collapse caused by climate change, while the economy continues to grow? An approach similar to that of venture capitalists would be ideal: many projects running for around 42 months and funded with a maximum of 2 million euros, well synchronized to coincide with three-year PhDs.

With Frameword Program 10, we will start an Artificial Intelligence revolution in science. For those who are about to shake up the European research landscape, we salute you.

*Afonso Ferreira is research director at CNRS, working at the Toulouse Institute of Computer Science Research.

Translation: Arthur Scavone.

Originally published in the magazine science/bsusiness

Translator's notes


[I] Um Frameword Program is a European Union research and innovation funding program that sets priorities and objectives for research and development over a given period of time, aiming to promote cooperation between EU member countries and encourage excellence in research and innovation across the Europe. (NT)

[ii] CNRS (Centre National de la Recherche Scientifique) is France's leading research organization. (NT)

[iii] Generative artificial intelligence refers to a branch of artificial intelligence that focuses on the ability to create new original data, content, or information. They are often used in creative tasks such as art generation, AI-assisted design, content creation, etc. (NT)

[iv] Google's AlphaFold is artificial intelligence software developed by DeepMind, an artificial intelligence technology company owned by Google's parent company Alphabet. AlphaFold was designed to predict the three-dimensional structure of proteins from their amino acid sequences. (NT)

[v] According to the company META, Calls is an accessible, open-source, broad-based language model designed for developers, researchers, and enterprises to responsibly build, experiment, and scale their generative artificial intelligence ideas. (NT)


the earth is round there is thanks to our readers and supporters.
Help us keep this idea going.
CONTRIBUTE