The Union of AI and Drug Discovery and Development Requires New Thinking for Structuring and Negotiating Strategic Transactions, and a Rigorous Analysis of Applicable Regulatory Considerations

The use of artificial intelligence and machine learning (AI/ML) in drug discovery and development is well established and here to stay. With AI tools developing the potential to impact virtually every stage of the life sciences product life cycle, the FDA continues to refine its framework for the use of AI to support regulatory decision-making. On April 7, the FDA ended its consultation period on two draft guidance documents, one containing recommendations on the use of AI to support FDA regulatory decision-making and another providing a “comprehensive approach” to the management of risk throughout the total product life cycle of AI-enabled devices. We also saw, in January of this year, the U.S. administration issue the Executive Order Removing Barriers to American Leadership in Artificial Intelligence. These developments, and potential developments, need to be carefully taken into account by life sciences companies that are considering transacting with an AI/ML tool provider.

We are continuing to see an increasing number of deals in which a drug developer partners with an AI/ML platform developer, and the synergies between these companies are immense, including at the development stage where it has the potential to accelerate drug discovery, expand the universe of therapeutic targets, and increase the efficacy of drugs against known targets.

However, both the new potential outcomes and the regulatory implications of these partnerships require the parties to think carefully about the optimal structure and terms for the transaction. The magnitude of the additional value from the use of AI/ML tools is not yet fully understood, and the manner in which these AI/ML technologies and the resulting therapeutics are treated from a regulatory standpoint is evolving and in its early stages.

While the FDA has been regulating AI for decades, this has historically involved technologies that were intended to be used as medical devices. Increasingly, pharma companies are using AI/ML not as medical devices but as tools to assist with product development or regulatory requirements, which poses some challenges to the FDA’s existing frameworks. Indeed, every individual use of AI/ML by a pharma company and an AI/ML company with which it wishes to transact is likely to come with its own distinct sets of benefits, risks, and challenges, and such use may or may not be subject to oversight from the FDA. The FDA regulates the use of AI throughout the product life cycle, but the level and nature of such regulation will differ significantly based on intended use.

In view of the above landscape, the following are important considerations when a pharma company and an AI/ML platform company are structuring and negotiating a transaction to license or acquire AI/ML for its deployment in drug development:

Milestone and Royalty Payments

The tried and true financial model consisting of an upfront payment, milestones reflecting success in clinical development and commercialization, and ongoing royalty payments is still the most oft-used approach for an AI/ML drug discovery collaboration. But companies may want to rethink the balance between these different levers. Is the success of the partnership driven by quantity – i.e., does the use of the AI/ML platform generate a larger volume of useful molecules or targets to potentially explore from a preclinical standpoint – and does it therefore warrant a larger upfront payment with less on the back end? Or is quality the focus – i.e., the molecules that are identified are lower in number but have a much higher chance of success in clinical development – and does this in turn lead to a lower upfront payment but higher milestones and royalties?

Nonexclusive License Agreements

Some AI/ML platform companies want to enable as many drug developers as possible to use their product, and as a result we are seeing a number of nonexclusive license agreements being signed. But drug developers may be a little wary of this type of arrangement, since it may require the input of highly sensitive data into the AI/ML product, the same product that one or more of their competitors may be using. AI/ML companies have – rightly and reliably – gone to great lengths to give assurances to licensees that their confidential information will be adequately protected. However, until these companies have a successful track record over many years, the newness of AI/ML products may still give some drug companies pause.

Exclusive License Agreements

One way to counteract this discomfort is to have an exclusive arrangement with the AI/ML platform company. As AI/ML products mature and some standouts emerge, we expect to see more acquisitions of these companies by drug developers. Until then, another possible option is a joint venture. Such a venture might enable the drug developer to obtain the exclusivity and control that it desires but would not foreclose the AI/ML company from partnering with other drug companies outside of the joint venture.

Regulatory-Related Deal Considerations

Generally, AI used just to identify potential candidates does not need FDA preapproval. However, it is vital for the acquirer or licensee to ensure a deep and rigorous comprehension of the use case of the AI and also to understand on which points the technology in question meets the regulatory criteria for being treated as AI. The ability to explain the AI’s role will likely be important during the FDA approval process.

In addition, the pharma company will need to identify and evaluate the impact of any existing rights – for example, privacy, consent, and licensing considerations – and to develop mitigation strategies for any risks that the AI/ML company has identified to date. It will also need to consider whether, and in what fashion, the partner has engaged with the FDA regarding its use of AI/ML technology, and – if the technology is being deployed globally – with other national applicable regulatory bodies, such as the European Medicines Agency (which applies the extensive EU AI Act).

The FDA’s regulation of AI will continue to evolve both with the use of new technology and the Trump administration, so it is important that pharma companies seek the advice of counsel to keep up to date with any significant changes. Similarly, modified and bespoke deal structures and terms will be needed to keep pace with technological advancements.

This post is as of the posting date stated above. Sidley Austin LLP assumes no duty to update this post or post about any subsequent developments having a bearing on this post.