Most of the time, human knowledge advances slowly. But every so often, there is a moment that revolutionises what we know.

ArchimedesAn ancient Greek mathematician from Syracuse on modern-day Sicily. discovered the principle of displacement. And last week, AI company DeepMindBritish AI company that is a subsidiary of Alphabet, Google's parent company. announced it had worked out the shape of almost every proteinLarge biomolecules made of up amino acids, which in turn are made up of DNA. known.

Proteins are the building blocks of life. Their shape is closely linked with its function. So if we can predict the shape of a protein, we can usually work out its function.

Mapping out these proteins was a colossal task for DeepMind's AlphaFoldAn AI programme created by DeepMind. It is a deep learning system, meaning that it can be taught how to draw patterns in a similar way to the human brain. AI programme. Previously, we had only mapped the structures of around one million proteins from 10,000 species. AlphaFold has now worked out the structures of 200 million proteins from about one million species.

It has also shown that it is capable of performing tasks that have stumped humans for years. One research team spent ten years trying to map the structure of a protein that can only be found in an ancient, extinct microbe. AlphaFold solved the problem in minutes.

Researchers have already used AlphaFold models to determine the structure of a protein that plays a key role in causing malariaA serious parasite infection transmitted by mosquito bites. It is one of the world's most lethal diseases.. Now they can use this to start developing vaccines.

Some proteins, known as enzymes, act as catalysts for chemical reactions. With some small modifications, we might be able to produce enzymes capable of digesting waste products.

This research could also do enormous damage. Protein-mapping AI could be used to produce biological weaponsA microbe or toxic substance that causes substantial harm to humans and can be used in warfare.. Enzymes could be programmed to be extremely harmful to humans.

Even if it is not used for weapons, some fear biotech could kill us by accident. Disease researchers often use biotech to create genetically-enhanced viruses, a procedure known as "gain-of-function researchDeliberately enhancing aspects, often deadly, of particular organisms. It helps predict emerging diseases and develop vaccines, but it has been controversial because of the risks involved.".

These viruses might be more virulent or more toxic. If they escape from the labs, they could cause catastrophic pandemics.

Another, less predictable danger comes from the study of mirror life.¹ Last year, scientists succeeded in producing mirror-image DNA that does not biodegrade.

This was a great step forward for science, but it comes with risk. Only some bacteria are capable of digesting mirror-image proteins. Producing too much could give them a huge evolutionary advantage.

Yes: While scientists are trying to use biotech for good, it can have too many unintended consequences. We need to slow the pace of discovery until we can work out how to contain the risks.

No: Every day, people are dying from diseases biotech could cure. It would be deeply irresponsible to slow the pace of discovery when so many people are desperately in need.

Or... You cannot put the genie back in the bottle. Biotech is getting cheaper and more advanced, and it would be futile to try to stop it or slow it down. Instead, we need to focus on putting together strong safety procedures.