A major Federal Circuit ruling just sent a clear message to AI-driven healthtech companies: AI alone won’t get you a patent.
In AliveCor, Inc. v. Apple Inc., the U.S. Court of Appeals for the Federal Circuit upheld the Patent Trial and Appeal Board’s (PTAB) decision invalidating AliveCor’s patents on AI-powered heart-rate monitoring. AliveCor had accused Apple of infringement via its Apple Watch, but the PTAB ruled the patents obvious over prior art—a conclusion the Federal Circuit affirmed.
The key issue? AliveCor’s patents applied machine learning to photoplethysmogram (PPG) data for arrhythmia detection. While prior art focused on ML techniques for ECG data or rule-based techniques on PPG-only data, the court nonetheless found it obvious to extend general ML techniques to the “precise use” of machine learning claimed in the patents, emphasizing that no claim required a specific ML algorithm.
This decision also undermines the ITC’s previous finding that Apple infringed, effectively blocking a proposed Apple Watch import ban.
The Takeaway for AI-Powered Healthtech?
- AI alone won’t make your invention patentable.
- Non-obviousness is key—generic AI+healthtech combo won’t cut it.
- Broad AI claims are risky. While patents that claim specific algorithms may suffer from their own risks (e.g., non-infringement and design arounds), broad claims covering general ML or AI techniques can be dismissed as obvious.
For AI-driven patents to stand, they must showcase a novel, non-obvious technological leap—not just AI applied in an expected way.