[FREE WEBINAR]Webinar replay, recorded May 26, 2023
Learn how to equip yourself with the knowledge and tools to protect your AI innovations and secure your position in the market.
[01]What's inside
The full webinar replay, free to request with the form.
Related services
Free download
Send the form and the replay plays on the next page.
An AI IP strategy decides how each part of an AI product is protected, so competitors cannot copy what makes it valuable. It matches each layer, from the model and the training data to the outputs and the product around them, with the right tool, whether that is a patent, a trade secret, copyright or a contract.
| Layer | Usually protected by | Why |
|---|---|---|
| A new model architecture or training method | Patent | Protects the technical method, even if a competitor writes it differently |
| Model weights and parameters | Trade secret | Kept confidential, they stay valuable without a fixed term |
| Training data | Trade secret and contracts | Licenses and confidentiality control how data is used and shared |
| The application built on the model | Patent and copyright | Patents for new technical features, copyright for the code |
| The product name and brand | Trademark | Protects the name customers know you by |
Patents and trade secrets pull in different directions, since a patent is published and a trade secret must stay hidden. See trade secrets for AI and patent eligibility for AI inventions.
Under current US law, only a human can be named as an inventor, and the Copyright Office registers only human authorship. AI-assisted inventions can still be patented when a person made a significant contribution, so record who contributed what as you build. Read can AI be an inventor.
Investors want to know why a better-funded competitor cannot copy you. Filed patents and well-kept trade secrets are concrete answers, and a clean chain of ownership avoids surprises in diligence. See why patents matter to investors.
Often you can patent a new technical method, such as a new architecture, training technique or way of applying a model to a specific problem. An abstract algorithm on its own is not patentable, and many companies keep the trained weights as a trade secret instead.
Usually both. Patent what competitors can see or reverse engineer from your product, and keep secret what they cannot, such as training data, weights and internal processes.
No. US patents must name a human inventor. AI-assisted inventions can be patented when a person made a significant contribution to the invention.
Send the short form. The replay plays on the next page, and you can come back to it any time.
[04]Learning Center
AI makers, whether building foundational models, AI agents, or AI wrappers, need patents, copyrights, trade secrets, and trademarks working together. Here is how each protects artificial intelligence, and where U.S. law stands on AI inventors and authors.
Can you patent AI? Learn what the USPTO requires, how the Alice test applies, and how to protect your AI invention. Get a free IP strategy call today.
AI training data acquisition involves critical intellectual property (IP) considerations that companies must address to mitigate legal risks and protect their…
[05]Free resources
GuidePDF
How to patent the AI features inside a SaaS product, written for founders.
Webinar replay
A webinar replay on using patents to build a moat around your SaaS features.
GuidePDF
How SaaS companies use patents to protect key features and add value.
Talk through your AI product with a registered patent attorney. The call is free, with no obligation.
