There is a lot of press around the use of artificial intelligence (AI), and the dangers it may pose. Some are concerned that AI may run amuck and destroy human society. Others are concerned that the technology could be used by bad actors to create propaganda that can be used to sway elections or topple governments. Like all technology, there is the risk that it will be misused but it also has the power to make serious, substantive, and positive impacts on our lives. One example is the impact AI is having on medical research. In March of this year, a couple of colleagues and I went to the Lab of the Future Congress in Boston and one of the big takeaways was the strides that have been made in using AI to speed up and streamline the process of identifying and vetting new drug candidates.

AI eliminates discovery lab work by using predictive methodologies that shorten the process of identifying potential therapeutic targets from months to weeks by analyzing large datasets, including genomic data, protein interactions, and disease databases. Machine learning algorithms can identify specific receptor targets for a drug, then perform in-silico synthesis of chemicals to match those receptor targets, more quickly unlocking the potential to assess and discover new pathways.

AI is used to further winnow these target compounds by simulating and predicting the efficacy of potential drug candidates. Machine learning models can analyze large libraries of compounds and predict their biological activity, transforming the process by which researchers narrow down the selection of promising candidates that can then go into pre-clinical studies. What historically took 4.5-5 years of time now takes ~2.5 years and at a reduced cost.

AI can be used in the lab to reduce the workload of bench scientists. Histopathologists spend 70% of their time identifying normal tissue on slides. AI is particularly good at identifying normal tissue slides, leaving a much smaller subset of abnormal slides to identify and categorize.

The promise of AI does not stop at the discovery phase, it also can improve and enhance the manufacturing process. The use of a digital twin, an in-silico model of an actual device or process, has been used to increase efficiencies, model the effects of scale up, and aid in technology transfers.

Throughout history new technological discoveries have been viewed in two lights. What positive aspects does it have, how can it help us,and how can it be abused? Technological innovations are not inherently good or bad, they just are. How they are viewed, as a positive or a negative, has more to do with how we perceive them to be used or abused, and less to do with the technologies themselves. The dangers of AI being abused, used for political purposes to sway elections, or even topple governments is just as valid as its promise for speeding discoveries, creating efficiencies, and saving time, effort, and money, in getting new drugs or vaccines to market. The difference is what it has always been. Us.

‍

‍