Data Science Investment Counter —Funding raised by UK data science companies in 2018.
£ 5.640 Million

BenevolentAI Teams up with Pharma Firm to Discover Drugs with AI

BenevolentAI has struck a deal with pharmaceutical firm Janssen to use Artificial Intelligence to evaluate new drugs.

The London-based company has developed deep learning algorithms able to discover and assess chemical compounds’ therapeutical properties by tapping into vast troves of data, studies and research publications.

According to the deal, Janssen licenced BenevolentAI to a certain number of clinical stage drug candidates; the company’s AI will analyse the molecules in order to understand whether they can treat certain diseases.

The deal grants BenevolentAI the right to develop and commercialise any medicine it were to discover during the process.

Founded in 2013 and based in the so-called King’s Cross Knowledge Quarter, the 60-employee company aims at harnessing AI to drive scientific progress.

That, so far, has mostly meant aiding researchers in discovering medicines , explains Jerome Pesenti, the CEO of BenevolentAI’s computer science branch BenevolentTech.

“The idea is to use all the data available out  there—publication, patterns, clinical data— to help scientists through the process of discovering new drugs,” he says.

Researchers using BenevolentAI’s system can simply interrogate it about a certain disease, triggering the AI to sift through the data and return a series of information on how the disease is currently treated. More importantly, the system also provides suggestions for possible lines of inquiry.

“Using Natural Language Processing, the system figures out connections— bits of information that can be found in those papers and publications, but have not been researched,” Pesenti says. “When someone is looking at a certain disease, the AI will come up with a list of targets— genes, compounds, molecules— that are worth looking closer at.”

Researchers’ feedback is eventually circled back into the system, making it smarter over time.

BenevolentAI’s computerised research assistant can also work out interactions and sides effects of a candidate substance, and recommend alternative molecules with similar properties.

“We are aiming for predictive biochemistry,” Pesenti says.

Pesenti, an AI guru who previously worked at IBM Watson, explains that what makes BenevolentAI special is that it is not exclusively focused on the computer science side— as most AI organisations, in his opinion, are.

“The point is that we put both AI and drug discovery scientists in the same room,” Pesenti explains.  “That’s very necessary today: given the maturity of AI, we need to pay attention to the feedback from users.”

The company has actually set-up a branch mainly focused on life sciences, BenevolentBio, which works in cooperation with  BenevolentTech.

This model could be replicated in the future as BenevolentAI sets its eyes on different scientific disciplines. What is next exactly Pesenti does not know, but veterinary drug discovery and even materials science could be interesting fields of application for the company’s AI.

Image via


Co-working space and blog dedicated to all things data science.

Subscribe to our newsletter