The global race for leadership in artificial intelligence (AI) is more contested than ever and, for now, there is no clear winner.
According to a recent report from Stanford University[1], last year US-based institutions produced 40 notable AI models, while China produced 15 and Europe, 3. A model is considered notable if it meets at least one of the following four criteria:
- It is presented in a scientific article with more than 1000 citations
- Get the state of the art in one benchmark recognized
- It has historical relevance
- Deployed in a significant use product[2].
Despite these production figures by the United States, it was discovered that Chinese models achieve near parity with American ones in two performance markers: which evaluates the knowledge and problem-solving ability of an AI (Massive Multitask Language Understanding), and the one that evaluates the code generation capacity of an AI (HumanEval).
These results come at a time when world leaders are stating that Winning the AI race is essential for national security and for advances in healthcare, business, and technology..
The United States was the country that invested the most in AI and carried out the most mergers and acquisitions in 2023. Currently, it continues to invest millions of dollars in an ambitious AI infrastructure project with the support of large companies such as OpenAI, Oracle and Softbank. However, it was this same administration that revoked an executive order from the Joe Biden government on safe and reliable AI.
At the same time, China has made AI a strategic priority, especially in talent training. The country has the largest number of AI patents in the world and wide access to data, the main resource for training increasingly powerful models, although without the privacy guarantees that other regions demand.
The situation is different in Europe, as it has more restricted access to venture capital funding and faces a talent shortage, in addition to its strict data protection rules. While this poses a challenge, it also represents an opportunity to lead a more ethical and transparent AI. In this context, If Europe manages to turn its regulatory approach into a competitive advantage, it could make a qualitative leap in the global race towards more responsible AI.
DeepSeek undoubtedly caused a frenzy in January of this year when it appeared on the scene with R1, a Chinese AI and chatbot model that, according to the company, is cheaper and offers comparable performance to that of the rival American model ChatGPT from OpenAI.
What has raised some concern among European authorities is that DeepSeek stores information on servers located in China, which could constitute a violation of the General Data Protection Regulation, which guarantees European citizens control over their personal data and the right to delete it from databases. If we add to this that China is among the countries with the greatest risk of cyberattacks, the general public is recommended to use this service with special caution.
AI models are becoming larger, more computationally demanding, and more expensive in energy terms, so applied research centers how i2CAT must focus on leading the development of sustainable solutions that make possible a responsible, efficient artificial intelligence that is in line with the challenges of the future.
In the coming years, AI will be a key enabler in our daily lives: from task automation, to decision-making support, including a tangible improvement in productivity. On the other hand, it will have a transformative role in the healthcare field, improving diagnosis, personalized care and the quality of life of a rapidly aging global population.
The race for AI is not just a technological competition. It is also a matter of vision, values and responsibility. And in this field, Europe—with talent, cooperation and purpose—still has much to contribute.
