The first drug created by AI to treat humans is now ready to be tested

The first drug created by AI to treat humans is now ready to be tested

British startup Exscientia and Japanese pharmaceutical company Sumitomo Dainippon Pharma claim to have developed the first drug created using artificial intelligence. DSP-1181 will be used to treat patients suffering from obsessive compulsive disorder. Clinical trials of the drug will start in March this year in Japan.

When developing DSP-1181, Centaur Chemist, Exscientia's AI platform, used algorithms to identify the chemical components that would best suit the treatment of patients, taking into account a vast database of parameters to be met. The research phase for a conventional medicine is around 4.5 years old, however, the companies indicate in a press release that they were able to complete it in less than 12 months.

Drug development process through AI Credits: Exscientia

If the first phase of clinical trials is successful, Exscientia plans to conduct tests on a global level. In addition to DSP-1181, the company is developing drugs to treat patients with cancer and cardiovascular diseases and hopes to have a new drug ready to test by the end of 2020.

In recent years, AI has revolutionized the health area, explained Paulo Novais, Full Professor at the School of Engineering at the University of Minho and outgoing President of the Portuguese Association for Artificial Intelligence (APPIA), and Lus Paulo Reis, Associate Professor from the Faculty of Engineering of the University of Porto, Director of LIACC / UP and President-elect of the Association, to SAPO TEK. Experts indicate that, in 2019, AI became decisive in this area, with an increase in intelligent systems that helped to improve analyzes, diagnoses and treatment recommendations.

Google, for example, is one of the technology companies that has been betting heavily on AI for the detection of diseases such as breast cancer. Early in 2020, the Mountain View giant announced that its deep learning system developed a model that managed to reduce the number of false positive cases by 5.7% and false negative cases by 9.4%, during a test phase in some North American hospital units.