April 25, 2025

Patents, Trade Secrets & AI: How to Legally Safeguard Innovation | Vincent Allen

In this episode of the An Hour of Innovation podcast, Vit Lyoshin sits down with Vincent Allen, managing partner at Carstens, Allen & Gourley, LLP, and a leading expert in intellectual property and technology law. Together, they explore how startups and tech innovators can protect their software, AI models, and inventions in today’s rapidly evolving landscape.
They dive into critical topics such as the differences between patents and trade secrets, when startups should start thinking about IP protection, and how open-source software impacts your ability to secure innovation. Vincent also sheds light on the legal risks of training AI models with public or copyrighted data, ownership rights over AI-generated content, and the growing concerns around AI hallucinations and misinformation.
Additionally, they discuss how startups can navigate global AI regulations, the competitive challenges posed by international players like DeepSeek, and practical steps to legally safeguard your company’s intellectual property both in the U.S. and internationally.
Whether you’re building software, developing AI products, or simply curious about the intersection of technology and law, this conversation offers actionable insights to help you protect your innovation and stay competitive in a complex world.

Show Notes

Vincent Allen is a managing partner at Carstens, Allen & Gourley, LLP, specializing in intellectual property and technology law. With a background in electrical engineering and decades of experience in patent, trademark, and trade secret protection, Vincent helps startups and tech companies safeguard their innovations. He’s also at the forefront of legal discussions around artificial intelligence, advising clients on the risks, compliance, and ownership challenges of AI in today’s rapidly evolving landscape.

Support This Podcast

Amazon - buy things that Vit Lyoshin recommends

https://www.amazon.com/shop/anhourofinnovationwithvitlyoshin

Webflow - Create custom, responsive websites without coding

https://try.webflow.com/0lse98neclhe

MeetGeek - Record, transcribe, summarize, and share insights from every meeting

https://get.meetgeek.ai/yjteozr4m6ln

Monkey Digital - Unbeatable SEO. Outrank your competitors

https://www.monkeydigital.org/?ref=102449

Podpage - Build a beautiful podcast website in 5 minutes (or less)

https://www.podpage.com/?via=vitaliy

Takeaways

Startups often wait too long to protect their intellectual property.
Proper agreements with founders and developers are crucial.
Trade secrets can be a viable option for protection if kept confidential.
Public disclosure starts a one-year clock for patent applications.
Patents last for 20 years from the filing date.
Open source software can complicate claims of proprietary rights.
Acting quickly is essential when trade secrets are compromised.
Ownership of AI models is tied to their creators.
Fair use in AI training is a complex legal issue.
AI-generated work raises questions about authorship and copyright.
The user of AI tools must ensure the output is appropriate.
Critical thinking is essential when using AI-generated information.
International regulations on AI vary significantly between regions.
Companies should implement policies to protect confidential information when using AI.
Disclaimers can help mitigate liability for AI-generated outputs.
Trade secrets are harder to enforce internationally than patents.
Engineers should understand the legal landscape surrounding AI.
Collaboration between engineers and legal experts is crucial for responsible AI development.