Artificial intelligence promises to shake up virtually every sector and industry. There are few applications it could be used for as important as human health - it is quite literally a matter of life and death.
However, the pharmaceutical industry is not known as a bastion of openness or innovation. The process of drug discovery has largely remained unchanged for decades. One company hopes to change this, dramatically accelerating the process using AI: BenevolentBio, a subsidiary of BenevolentAI. Techworld spoke to CEO Dr Jackie Hunter.
What is BenevolentAI's main mission?
BenevolentAI’s mission is to apply its AI technology to accelerate scientific discovery and address some of humanities’ most pressing challenges, with human health being a starting point.
Healthcare is an area where applied AI can have a tremendous effect, helping to revolutionise the drug discovery process by creating a new scientific method, one led by technology. Currently the process is hugely expensive, subject to high failure rates and it takes too long to bring a drug to market.
The problem is that despite the huge growth in scientific knowledge, this process of discovery hasn’t changed for half a century. As such, it’s impossible for humans to process all the world’s scientific information and turn into something tangible. BenevolentAI is combining the power of AI with the expert insight of its scientists to help accelerate the drug discovery process, dramatically.
What are the products/services BenevolentAI offers?
We have developed a unique AI platform that helps scientists to make new discoveries. This platform generates functional knowledge from large amounts of unstructured and semi-structured scientific data found in patents, biochemical knowledge, scientific papers and clinical trials.
Our technology uses Deep Learning and Natural Language Processing to analyse, normalise and understand language to build reliable knowledge. Algorithms then explore that knowledge and learn and reason to create unique hypotheses. The platform then helps scientists and researchers to validate these hypotheses. This insight can then be fed back into its machine learning models to help them improve and learn for the next process.
BenevolentAI’s technology facilitates previously impossible scientific discoveries by finding connections that might not have been found by human brain power alone. The first use case has been applying our software to automate the mining of biological knowledge, undertake predictive and generative biochemistry, and advance better therapeutic molecules. This helps our drug discovery scientists to create new medicines more efficiently.
Who are your customers?
The business model is to create revenues from the assets that the technology creates, not the technology itself. For example, where scientific discovery leads to new medicines, revenues are generated from selling the patent rights to these molecules, partnering with pharmaceutical companies to take these drugs to market, or fully developing the drugs ourselves and marketing those medicines directly to patients.
What are the main breakthroughs within AI recently?
AI can not only innovate the existing drug discovery process in terms of cost and efficiency. Perhaps the most exciting application of AI is that it enables previously impossible research and can deliver previously undiscoverable cures and treatments for diseases. The company believes that its recent deal with Janssen Pharmaceutica NV (a license for a series of novel drug stage candidates) marks the start of an era where AI technology and the traditional healthcare industry will work together to improve and innovate the drug development process.
What are the main benefits of BenevolentAI's tech, and indeed AI/machine learning generally?
According to Eroom’s law, the cost of drug development doubles every nine years. In a sector where the drug discovery process hasn’t changed in decades and is highly inefficient; time and costs are increasingly valuable commodities. AI enables intelligent mass analysis of large amounts of scientific information to create new, usable and deep knowledge which significantly accelerates the process we’re performing.
It greatly reduces the margin of error when new drugs are put into clinical trials, helping to reduce the vast resources and money that are needed to deliver fit for purpose drugs. This is just one example of how we’re using AI in healthcare, but the same process can be transferred across a range of sectors.
What are the risks you foresee with AI? How do we combat them?
The company believes AI technology and machine learning can be harnessed for ‘benevolent’ ends. We feel that AI and machine learning do not supplant human intelligence but instead augment it and such augmentation is essential in solving some of the world’s greatest scientific challenges. In order to ensure AI follows the morals and ethics that we hold ourselves to, we must integrate them into the technology as we develop it.
The real key to ensuring AI does what we want it to do, is through the information we feed it. If you use inaccurate, skewed or harmful information, logic suggests that you’ll create AI that has certain moral glitches and predisposed biases – those not necessarily in-line with society’s ethics.
When it comes to BenevolentAI’s use of AI self-learning, it aims to feed its systems with information that will truly benefit society. In particular, information that can be used to treat, solve and help us to learn more about the things that threaten and hinder society.
Whether it is looking for the cures and better treatments for diseases, improving energy efficiency or finding ways of cutting carbon outputs. There really is no end to the enormous benefits AI can deliver. As long as we ensure the information and data the complex algorithms that is used comes from trusted sources.
What are your plans for 2017?
2017 will see BenevolentAI continue to build and develop its drug development pipeline and includes out first PhaseIIb clinical trial. This year we will also begin to expand the use of our technology beyond bioscience to other scientific industries.
What new tech can we expect to see - what excites you most?
Currently we are focusing the application of our technology on human health, but are beginning to seek ways to expand into other scientific industries such as veterinary medicine, nutraceuticals, materials science, energy and agriculture.
The possibilities that our technology enables are endless but we want to manage growth and market focus carefully. AI is perhaps the most exciting area of technology at the moment and we can only see that growing as it improves and advances across a multitude of sectors to benefit society.