The artificial intelligence (AI) revolution is here, and it's going to upend the biopharma sector in ways that most investors might not have considered. The end results are impossible to predict with accuracy, but based on what's possible with the AI tools of today, investors could reap tremendous returns from companies that are aggressive with implementing AI into their workflows.
Specifically, there are at least three ways that AI is going to disrupt biotech stocks in the near term. Let's take a look at them.
Image source: Getty Images.
1. Drug development time could shrink -- and success rates may rise
It takes an average of around 10 years and $2.6 billion to move a candidate from pre-clinical testing through clinical trials and the regulatory approvals to reach the market, so there's a lot of room for improvement. The most important way that AI will disrupt the biotech industry is by compressing the time it takes to develop a new medicine and get it approved for sale, and a few players are already working to realize that benefit.
Consumer genetics and biotech company 23andMe (ME -2.89%) is already using machine learning to help its collaborators, specifically GSK, approach the process more effectively from the earliest phases of development. With its technology, 23andMe claims that it can cut the time before a company can move from pre-clinical testing to clinical trials with one of its candidates from an average of 7 years to roughly 4.5 years.
In due time, using AI to accomplish similar reductions will become the norm. But for now, the early movers will experience the biggest advantage, and they'll get their medicines out the door much faster as a result.
Likewise, it's probable that AI will help to reduce the risk of companies failing in their clinical trials by helping to screen out the weaker, early-stage programs based on the totality of what's documented in the relevant scientific literature. 23andMe is already suggesting that its approach might lower failure rates though it lacks concrete evidence as of yet.
Still, with only 13.8% of programs managing to make it from phase 1 trials to approval for sale and reaching the market, there could be massive benefits to the biopharma businesses that implement such AI solutions -- and at some point, those benefits will probably be realized.
2. Clinical trial costs could fall sharply
A significant portion of the costs involved in running a clinical trial are for employing clinical coordinators and other staff whose jobs involve screening, recruiting, orienting, and otherwise managing the study participants to ensure that they show up at the right place and the right time to get treated with the investigational medicines.
On average, each patient in a trial costs about $36,500, though that sum can vary by tens of thousands of dollars depending on the type of intervention and the target indication, among other factors. If even some of that (massive) workload and cost burden can be offloaded onto artificial intelligence systems, it will significantly reduce overhead.
Such an AI system need not be very advanced, and it's almost a slam dunk that existing models like OpenAI's GPT-4 could be adapted for the task. For example, rather than needing to talk to a person, clinical trials could easily use coordination chatbots capable of scheduling people for study visits, answering most of their questions about how to prepare for their appointment, and checking in with them as part of follow-ups to survey their side effects.
Such a system need not be only text-based. With real-time voice synthesis and advances in audio processing, patients could even talk in natural language to the AI coordinator over the phone.
But it's unlikely that human clinical-research coordinators are going to be unemployed anytime soon, as they fulfill so many different roles. Instead, clinical-trial sizes could balloon as the marginal costs of including additional patients will be significantly lower.
That could lead to higher-quality data, and it could also mean that smaller biotechs might be able to run trials that right now only large companies have the resources to do. As the costs fall, big participants like AbbVie could potentially pass the benefits of lower costs on to their investors via increases to the dividend.
3. Commercialization costs are also on the chopping block
Selling, general, and administrative (SG&A) costs tend to stay low for biotech companies until they get a stamp of approval from regulators and it's time to try to market their medicines. AI could significantly drive down many of the costs associated with marketing, like developing graphics, writing advertising copy for the website, and making brochures for prescribing physicians.
Right now, AI image-generation models like StableDiffusion can make photo-realistic pictures of products, complete with smiling patients experiencing newly improved health outcomes. Similarly, text generation for airy marketing literature is already within the purview of existing and publicly usable AI models.
This isn't science fiction -- these are cost savings that are waiting in the wings for any biotech willing to chase them. And those that move first will be able to secure the most gains for their investors.
