2025 and the new frontiers of medicine thanks to AI | Google ai | Non-generative ai examples | Generative ai financial services | Turtles AI

2025 and the new frontiers of medicine thanks to AI
New Perspectives for Pharmaceutical Research Thanks to AI
Isabella V22 January 2025

 


The year 2025 could mark a momentous turning point for the pharmaceutical industry: according to Google DeepMind CEO Demis Hassabis, the first drugs designed with the help of AI will be ready for clinical trials. During a speech at the World Economic Forum in Davos, Hassabis revealed that by the end of the year, the first trials are expected to begin on drugs developed through advanced machine learning techniques, the main goal of Isomorphic Labs, a DeepMind spin-off dedicated to pharmaceutical research.

Key points:

  • Accelerated timelines: AI could significantly reduce the time needed to develop new drugs, currently estimated at 12 to 15 years.
  • Reduced costs: AI-based solutions could lower development costs, which average $2.6 billion per drug.
  • Challenges in data: Lack of high-quality data is an obstacle, which can be mitigated through synthetic data generation and clinical collaborations.
  • Collaboration and innovation: Major technology and pharmaceutical companies, including Nvidia, are investing in the potential of AI for medicine.


Enthusiasm for AI’s capabilities in improving medicine is supported by tangible results such as AlphaFold, technology capable of accurately predicting protein structure, a key advance in understanding and developing new treatments. This work has already earned Hassabis and colleague John Jumper the Nobel Prize. However, Hassabis emphasizes how AI, while capable of solving complex problems, cannot yet replace human creativity in formulating scientific hypotheses.

One of the main advantages of introducing AI into pharmaceutical research lies in the possibility of optimizing personalized medicine by quickly adapting it to individual metabolic specificities. This approach, according to Hassabis, could drastically reduce development time, with prospects for real improvement in the traditionally low success rates of clinical trials: only 10 percent of these lead to actual drug approval.

Despite the promise, the industry faces challenges related to the availability of high-quality data, hampered by privacy issues and acquisition costs. To overcome these barriers, synthetic data are being used, while recognizing the risks of possible bias in their use. Hassabis insists on the importance of generating targeted data and collaborating with research institutions to fill any gaps.

In addition to Google DeepMind, other companies are embracing AI for drug discovery. Nvidia, for example, has made BioNeMo, a set of tools for molecular machine learning, open source and is collaborating with pharmaceutical giants such as Novo Nordisk to develop advanced supercomputer-based research systems such as Gefion. These initiatives highlight the growing interest in the integration of AI technologies and biology, paving the way for a new era of medicine.

The advance of AI is transforming the landscape of scientific research, propelling us toward increasingly innovative and accessible medicine.