Thanks to artificial intelligence, a new super antibiotic has been identified: called 'halicin' in homage to the Hal 9000 supercomputer of the film '2001: A Space Odyssey', it has shown in laboratory tests that it can eliminate many of the disease-carrying bacteria, including some strains that have become resistant to all traditional medicines. The result, published in the journal Cell by researchers from the Massachusetts Institute of Technology (Mit), could represent the beginning of a new era for antibiotic research, which has slowed sharply in recent years.
"We are facing a growing crisis as regards antibiotic resistance, and this situation has arisen both due to an increase in pathogenic microorganisms that have become resistant, and to an increasingly scarce production of new antibiotics by biotech industries. and pharmaceuticals, "says James Collins of Mit.
To overcome this impasse, researchers have developed a machine learning algorithm trained to analyze extremely accurately the chemical structure of chemical compounds, correlating it with specific properties such as the ability to kill bacteria.
In this way, they obtained a platform capable of reviewing over a million million compounds over a few days, identifying those that could kill bacteria using different mechanisms of action compared to existing drugs. "Our approach has unveiled this fantastic molecule which is arguably one of the most potent antibiotics ever discovered," stresses Collins.
The study also identified 8 other potentially interesting molecules that will be tested shortly. According to the researchers, the system could also be used to design new drugs and optimize existing ones.
High-risk artificial intelligence (AI) systems, such as facial recognition, will have to be "transparent, traceable and guarantee human control" in 'sensitive' sectors such as "health, police and transport" : so the EU Commission in its White AI paper presented in Brussels, adding that "these systems can pose risks" for this "it is essential to build trust" with "clear rules" for "high risk" applications that must be "compliant with European standards".
All artificial intelligence applications arriving on the EU market "are welcome" but will have to comply with the European regulatory framework, warns Brussels. For low risk systems, the Commission provides for a voluntary labeling system . The European Commission also clarifies that the authorities "should be able to test and certify the data used by the algorithms", guaranteeing "respect for fundamental rights, in particular non-discrimination".
In the 'White Paper' on artificial intelligence, the EU Commission also stressed that facial recognition is among the high-risk artificial intelligence systems, "its use is generally prohibited" and allowed "only in exceptional cases, duly justified and proportionate, subject to guarantees and based on Union or national law ". The EU executive intends to start "a broad debate on what possible circumstances could justify these exceptions".
"Artificial intelligence is not good or bad in itself: it all depends on why and how it is used. We allow the best possible use and control the risks that artificial intelligence can represent for our values - no harm, no discrimination! ". So on Twitter the vice-president of the EU Commission with responsibility for digital and competition, Margrethe Vestager .
To strengthen artificial intelligence (AI), the goal is to attract investments of € 20 billion per year over the next 10 years: the EU Commission writes in its communication today on artificial intelligence, data and digital.