AI: Speeding Up Science, Changing Healthcare
AI tools fast-track scientific breakthroughs and reshape medical care, like drug discovery and diagnosis, but raise ethical concerns.
Topics: science, technology
Example
In July 2022, Google DeepMind and EMBL-EBI released AlphaFold's predicted structures for over 200 million proteins, nearly every protein known to science, and made them free to use. Knowing a protein's shape helps scientists understand diseases and design drugs. In 2024, DeepMind's Demis Hassabis and John Jumper shared the Nobel Prize in Chemistry (with David Baker) for this protein structure prediction work.
Key facts
- In July 2022 Google DeepMind and EMBL-EBI released AlphaFold's predicted structures for over 200 million proteins, free to use.
- Demis Hassabis and John Jumper shared the 2024 Nobel Prize in Chemistry for protein structure prediction, with David Baker.
Limitations
- AlphaFold predicts shapes; it does not itself produce medicines, and predicted structures still need lab checking, so claims that it 'cured' diseases go beyond the evidence.
Evaluations
science evaluation
Support
AlphaFold shows AI solving a 50-year-old scientific problem: predicting a protein's 3D shape from its sequence, work that once took years of lab effort per protein.
Counterargument
Predictions are less reliable for flexible proteins and for how proteins interact with drugs, so researchers still rely on experiments for the details that matter most in medicine.
Rebuttal
Even with those limits, starting from a good predicted structure saves time on the many proteins nobody had studied, which is why the Nobel committee honoured it.
Additional support
Releasing the database free through EMBL-EBI let labs in poorer countries use the same structures as wealthy institutions, widening who can do this research.
technology evaluation
Support
The 2024 Nobel Prize recognised AI work by a technology company, a sign that machine learning has become a core tool of science rather than a side project.
Counterargument
Building such systems needs data and computing power that only a few large firms command, which concentrates control over future scientific tools.
Rebuttal
DeepMind's choice to release the 200 million structures openly shows that concentration is not inevitable; it depends on how firms and funders choose to share results.
Additional support
AlphaFold was trained on decades of structures that publicly funded scientists had deposited in open databases, so it built on shared human work rather than replacing it.
Sources
- Google DeepMind (2022-07-28): AlphaFold reveals the structure of the protein universe
- Nobel Prize Outreach (2024-10-09): The Nobel Prize in Chemistry 2024 - Summary