Artificial intelligence is transforming how new therapies are discovered, developed, and delivered, but for early-stage biotech companies, that transformation comes with a minefield of legal risk.
At LaunchBio’s recent NextGen VC Forum, one of the day’s most talked-about panels tackled the murky intersection of AI, intellectual property, and biotech innovation. Moderated by leading attorneys from Wilson Sonsini, the conversation explored what biotech founders, investors, and commercial partners need to understand about the legal frameworks surrounding AI.
Trade Secrets Over Patents: A Possible Shift in Strategy for Platform Protection
Biotech has traditionally used patents for protection of its intellectual property and proprietary platforms, but AI companies are considering a different approach for protection of their AI platforms. Wilson Sonsini speakers noted that today algorithms can be replicated across languages and frameworks, and enforcing software patents is often impractical. That’s why many companies — including the biggest players — are turning to trade secrets instead.
This approach isn’t foolproof, though. The portability of talent poses a significant risk. When engineers can re-create core functionality from memory at a new company, the boundaries around proprietary tech start to blur. Protecting against leakage now requires more than just legal paperwork, but it also requires smart internal governance and culture-building.
Data Provenance Is the New IP Diligence
Perhaps the most urgent legal risk raised was training data. As generative AI tools proliferate, so do questions about where the underlying data came from, whether it was scraped legally, and whether its use aligns with terms of service, privacy laws, or even international regulations.
What used to be an afterthought is now a central diligence item. If a biotech company is using AI to drive R&D or commercialization, investors and partners are beginning to ask: Do you own your data? Do you have the rights to use it? Can you prove it? The consequences of getting it wrong are serious. One speaker shared that improperly sourced training data can require a full rebuild of a model, a costly and time-consuming setback for any startup.
AI-Assisted Inventions and the Humans Behind Them
The panel also highlighted a growing gray area: patenting inventions developed with the help of AI. It’s no longer a hypothetical. Regulatory and legal challenges are emerging around whether certain inventions count as human-made, particularly in fields where generative tools are used to identify new compounds or molecular pathways.
Documentation is key. Founders are advised to track human contribution in the invention process now, so they aren’t caught off guard later, especially in the face of future challenges from generic competitors or regulatory bodies.
Why It Matters Now
From contracting to compliance, the discussion made it clear that AI isn’t just a tech issue, it’s a governance issue. For life science companies, especially those handling patient or health data, internal policies around AI usage are quickly becoming a best practice, not a nice-to-have.
The AI panel was one part of our full, half-day NextGen VC Forum. It featured new sessions on national security laws and cross-border dealmaking, all curated to help the VC and biotech communities think more proactively in a landscape that rewards agility.
Want in on the next one?
The fall 2025 NextGen VC Forum takes place in Boston on October 21. With limited seating and an invitation-only guest list, this is the place to deepen your understanding of the life science and biotech sectors while building meaningful connections across the industry.