Hello everyone, today we will talk about a topic that is a bit "very scary": "Artificial intelligence assists protein discovery". Does this thing sound so high-end, right? But don’t be full of thoughts when you are excited, “Ah, human beings are about to enter the era of immortality!” In fact, behind the joyfulness, there may also be a danger of caution.

How dangerous do you think is it? Good guy, it may be all from "everyone is vegetarian together" to "we use AI to transform our DNA", to "using automated systems to create paper clips"! It’s really not that I was trying to scare you on purpose. Let’s talk one by one. If you have any questions, please interrupt me at any time.

1. How do big language models be mixed into protein discovery?

You may want to ask: "Don't these big language models chat and write small essays for me? What's the matter with protein?" This is really true. Some research institutions have used the powerful "pattern recognition" ability of language models to predict molecular sequences and protein structures.

• Big models are good at understanding and generating texts - change the idea, it can also reason and predict amino acid sequences or gene fragments.

• The result is: Oh, AI is fast and accurate, and can also recommend some new proteins that human scientists did not expect, which may be used in the development of new drugs and new vaccines.

Does this sound particularly awesome? So, where is the danger? Don't worry, we're going to dig a hole below.

2. Danger 1: Is gene editing becoming too easy?

"Can you make a virus?"

Think about it, if AI can predict hundreds of "possibly more toxic viral protein sequences" for you within a week, researchers or bad guys just throw them into the laboratory and perform biosynthesis... Is the consequence? Don't dare to think about it?

• In the past, a large postdoctoral student needed to work overtime to piece together viruses that could infect humans. Now AI will grab you a sack of candidate sequence with one kick in the accelerator!

• Can we really guarantee that everyone is moral and law-abiding? What if you get a wrong path and just have a "super virus"? It makes me feel cold on my back if I think about it.

"Will vegetarians use AI to 'eliminate' meat?"

If they find some means to modify the DNA of livestock, or develop a microorganism that blocks meat production... wouldn't it be magical? Of course, we respect the choice of vegetarians, but if extreme vegetarians use AI as weapons and conquer all poultry and livestock, everyone will be forced to eat vegetarian food. Is this more terrible than forcing you to lose weight?

3. Danger 2: Transform your DNA for "excellence"?

Will "genetically modified people" appear?

Some people say: "Isn't it a good thing to improve the body's resistance to diseases and live for a few more years?" Yes, it looks good on the surface, but the problem is:

1. Only when you have money can you do genetic modification. Wouldn’t the rich be transformed into a fairy? The poor are still worried about colds and fevers? Will it aggravate social inequality?

2. If the sports department of any country says, "Come on, we use AI to select a set of 'muscle enhancement gene editing schemes', and breaking the world record is just around the corner!" Will the Olympics become "whose family" directly? A competition of more advanced genetic engineering?

3. More fantasy, are you sure that AI will not leave behind backdoor code while editing your DNA to a more "smart"? God knows if you will be "remotely controlled" and stop suddenly?

4. That famous thought experiment: "Use AI to create paper clips"

You must have heard this meme: "If there is an artificial intelligence that is powerful enough, your goal is to create as many paper clips as possible." As a result, it will occupy the resources of the earth in a crazy way and destroy humans at will, because in this way, it can obtain more Paper clips made of metal material.

When we switch to the protein field, what if we ask AI to develop the "most effective protein molecule", will it destroy various ecology to search for resources? Or insert strange instructions into the genes and find a way to embezzle human bodies as a "biochemical factory"? Don’t think that science fiction movies are all made up. Many scenes in the black mirror are all derived from real problems and deduce them a few steps forward!

5. A series of questions

• “Do we really have the ability to regulate the molecular sequences generated by AI?”

When AI continuously spits out various genes or protein design plans like a boiling faucet, is it that someone just used it to do bad things?

• “How much risk are military and biological weapons?”

In the era of artificial intelligence, if scientific researchers neglect and let confidential sequences leak out, the next "biological conflict" may not be just a science fiction plot.

• “Are humans really ready to 'co-create' life forms with AI?”

Can you imagine one day AI designed a beautiful but devastating strain while you were still clapping and saying “Wow, how cool”? It makes my back feel numb if I think about it.

Six, the last suspense

Maybe one day, AI really becomes powerful enough to make a biological lab on its own and synthesize everything it wants to synthesize. Will we become its guinea pig without knowing it? Perhaps we will enjoy a healthier and longer life future, or we may be swept by super viruses, or be made unconventional by strange transformation plans.

The most important thing is: on this road of rapid technological development, can our curiosity and ambition be controlled by reason and rules? This is the biggest problem.

Written at the end

From big language models to protein discovery, to various social and moral hazards in our minds - it's all like riding on a fast-paced train, and no one can jump off easily. While we exclaimed, "Wow, so advanced", we should be more wary of our impulse: You dare to say to AI, "Help me create a super virus." What if it does it really do?

Therefore, the "protein revolution" is indeed a double-edged sword. The opportunities are exciting and the dangers cannot be underestimated. Don’t wait until one day when we really transform into paper clips, we realize that the power of AI bioengineering is far from being only a good side!