In an interview, the chemistry Nobel Demis Hassabis suggested that artificial intelligence will soon accelerate the development of medicines and even healing. Experts examine the controversial declaration. Ten years, the Nobel Prize for the Chimica Demis Hassabis aims to treat all diseases with the help of artificial intelligence (IA) -Aen this is what has been read on the internet. The CO -Fondor and executive director of Google Deepmind specialized in AI, was interviewed on various aspects of this technology in the 60 -minute program from CBS News.
Together with his colleague John M. Jumper, a 48 -year -old Englishman developed in Deepmind the Alphafold2 model, able to predict the structures of practically all 200 million proteins now known. These substances perform a series of biological functions and ailments in their production, structure or function can cause pathologies. The discovery was worth the duo The Chemistry Award in 2024.
In an interview with 60 minutes, Hassabis hypothesized that in the future the IA will be able to shorten the development of medicines in weeks or even months. He added: “And I think that one day we could be able to heal all diseases with the help of the AI”.
The interviewer Scott Pelley asked: “The end of all diseases?” This is at hand, Hassabis confirmed: “Maybe even by the next decade, I don’t see why.”
Revealing molecular structures
Sometimes it is possible to estimate the function of a particular protein from its three-dimensional structure, but “we do not know that in most proteins in the human body”, explains the biochemistry and computer scientist Katharina Zweig, director of the algorithm liability laboratory at the Kaiserslationn-Clandern-Technical University in Germany.
The change in the protein structure is the cause of some diseases and “therefore it is possible to develop a medicine to prevent this process”. First, it took an entire doctoral thesis, requesting from three to five years to identify a single molecular structure, calculate and shape it. In this sense, “the age of Hassabis is, in fact, a revolution”, confirms Zweig.
Florian Geissler, head of the Fraunhofer Institute of Cognitive Systems (IKS) researcher, recalls that the causes of the disease usually cannot be reduced to an element “, but there are many examples in which proteins play an important role”. So, “in the coming years I will allow us things that we cannot even imagine today”.
However, Katharina Zweig ensures that all diseases within ten years will not yet be treated, since absolute correspondence has not been established among the many proteins and some diseases: “There are mutations with three abnormal -dimensional structures. Statistically it seems to be the cause of pathological symptoms, but in reality they are null”.
And even if it is defined which protein structure it appears in which disease, a long process is needed until a drug enters the market: “It is necessary to test clinical studies, which require enough patients, authorizations. So I think it will not be so fast”.
Current uses of artificial intelligence in medicine
In the diagnosis based on images of computerized tomography, the IA recognizes the presence of pathological changes much faster, explains Florian Geissler. The technique also helps in case of unexpected side effects in a combination of medicines, allowing you to optimize treatment methods.
Artificial intelligence also has the potential to reduce the load on the health system “, for example by automatically summarizing conversations with patients and preparing structured relationships for health boxes”, says the researcher. “This allows you to save precious time on the health system. Here the IA will play a decisive role.”
Zweig biochemistry, however, that, despite the assistance to artificial intelligence, the care of diseases requires great financial resources: “therefore the medicines will continue to be developed only where there are enough patients with money and then pay them”.
Diagnosis of limited value
Another question is that, in general, it is rarely possible to pronounce a diagnosis based on defined rules, adds Katharina Zweig. One of these cases is diabetes, in which “there is a limitation rate and clear measurement methods”.
Most of the other diagnoses, however, require much judgment and experience and “I do not know any artificial intelligence system today which makes them so reliable to the point of replacing doctors and doctors”. Florian Geissler begins equally from the principle that, in the near future, the decision on the treatment will remain human responsibility, “especially for ethical and legal reasons”.
And perhaps also because, at the moment, artificial intelligence systems are still a sort of “black box” “, in which an input occurs and receives an answer, but unconsciously one hundred as the decision has been made”, compares the researcher of the Fraunhofer Institute.
Katharina Zweig describes this: “We have no way to see automatic learning, nor how it is your diagnosis. Therefore we cannot conclude if we start from criteria that we, as human beings, would also have established”.
Source: Terra
Rose James is a Gossipify movie and series reviewer known for her in-depth analysis and unique perspective on the latest releases. With a background in film studies, she provides engaging and informative reviews, and keeps readers up to date with industry trends and emerging talents.
