Microsoft's Cell2Sentence-Scale Model Identifies Silmitasertib as a CK2 Inhibitor That Increases Antigen Presentation in "Cold" Cancer Cells – A New Insight Tested in the Lab and Demonstrates AI Research Creativity
Not often Microsoft CEO tweets enthusiastically About a scientific breakthrough in cancer research. But it can be understood when Microsoft's invention is the one that made that same breakthrough – and in the same breath also proved that artificial intelligence can reach scientific discoveries on its own.
To understand what AI has done, we need to first review one of the biggest challenges in fighting cancer. Many tumors are 'cold', meaning they are 'invisible' to the body's immune system. As a result, white blood cells are unable to recognize the cancer cells as such, and ignore them completely.
Microsoft researchers have developed an artificial intelligence system called Cell2Sentence-Scale. You can think of it as GPT-chat, but in the language of cells and the molecules they are made of. They gave this artificial intelligence a task: find a substance that would cause cancer cells to signal to the immune system that there was a threat to the body. The machine went through more than 4,000 different substances, and identified several that were supposed to be able to do the job.
Did she identify the right substances? About a third of them were already known in the scientific literature on the subject. But the other substances surprised the researchers because they had not been linked to these kinds of effects before.
One of these new substances identified by the system is called silmitserative, which in turn suppresses another molecule in the cell called CK2. Artificial Intelligence She predicted That this suppression will cause cancer cells to release signals that will cause the immune system to target them.
“What made this prediction so exciting was that it was a completely new idea,” the researchers wrote. “Although CK2 has already been shown to be involved in many functions in the cell… its suppression by silmitsertiv has not been explicitly reported in the scientific literature to enhance antigen presentation [which causes the immune system to target cancer cells]… This demonstrates that the model generates a new, testable hypothesis, rather than simply restating known facts.”
Well, the AI made a prediction, but was it correct? The researchers took silmitsertiv to the lab and tested it on normal human cells, as well as those that had been tricked into behaving like cancer cells. They saw that this molecule had no effect on the normal cells. But on the ones that had behaved like cancer cells? On them, silmitsertiv worked like magic, and they tried to report to the body's immune system the threat they posed.
what is the meaning?
First of all, we can stop with the absurd claims that "artificial intelligence has no imagination or creativity, because it can only recycle previous ideas." This logic has been problematic since the early days of the great language engines, because they did not copy entire sentences, but rather recombined words according to probabilities. And if those words led to the creation of new ideas - how good! Every new idea is the result of existing words that are connected in new and interesting ways.
This research demonstrates that artificial intelligence can reassemble ‘words’ – or proteins, or genes, or cells, or any other concept – in a creative, logical, and meaningful way. Yes, it has creativity. No, it doesn’t just ‘recycle.’ We can move forward.
Second, we see here how artificial intelligence can advance medical science. Or any science in general. The system demonstrated in this study Open for general useEvery researcher Can run it And start testing ideas and asking it to formulate new hypotheses of its own. Hell, researchers, leave it alone. Any kid can use it, and that's probably what's going to happen. Don't be surprised if in the next few years, some of the new drugs come from smart, bold teenagers. They'll use systems like this to come up with ideas for new drugs, and then they'll find a university professor who'll accept them into his lab and let them run experiments to prove their claims. And from there, it's a short walk to clinical trials.
Artificial intelligence will run itself
Or, of course, artificial intelligence will run itself, alongside automated experiments in robotic laboratories, which will lead to the generation of new drug ideas and their automated testing at previously unknown rates.
If that's what happens, we'll enter a new era of science. A time when new drugs, innovative ideas, and creative theories will emerge at an incredible speed. The main bottleneck will be testing these ideas at the same rate they are produced. But with the help of robotic laboratories, which will conduct experiments more efficiently than any human researcher, we could also accelerate the pace of scientific research itself.
Last but not least, as I have written several times before, we are entering a particularly exciting period in the history of science and technology. If the pace of development, understanding, and invention increases as I suggest, then it will be harder than ever to predict the future. Even so, it is difficult for us to predict how we will use innovative technologies. Who would have thought, for example, that the main use of GPT chat would be as a psychologist for humans? Or that people would use artificial intelligence engines to produce videos of animals jumping on trampolines at night? Now think that in a year we will have another revolutionary technology that will change society once again, and after that another, and another and another. How can we seriously talk about the year 2030, when the big changes just keep coming and piling up on each other, year after year?
But these are the troubles of futurists. For everyone else, this rapid progress will certainly cause "future shock" and the feeling that everything we know is being undermined, but alongside these it could also bring dramatic improvements in living conditions and health.
What would you prefer? A slow, calm, disease-filled life, or a frenetic, hectic existence that leads to vast improvements in our health—and perhaps even the promised eternal youth?
We may have to decide in the coming decades. And maybe even sooner than we think.
More of the topic in Hayadan:
