Historical Context of Patent Eligibility

In 1980, the U.S. Supreme Court’s ruling in Diamond v. Chakrabarty established the principle that Congress intended for “anything under the sun that is made by man” to be patentable. This landmark decision set a broad precedent for patent eligibility, embracing various inventions.

The Shift Introduced by Alice

The legal landscape shifted dramatically in 2014 with the Supreme Court’s decision in Alice Corp. v. CLS Bank International. This ruling introduced a new standard for assessing whether business method technologies were patentable, hinging on whether the claims were considered “abstract ideas.” Such an evaluation often led to additional scrutiny, limiting the scope of what could be patented, especially in the realm of software.

The irony of this development lies in the essence of patents, which distill an invention into its fundamental components—an abstraction, paradoxically deemed non-patentable if it resembles mere office procedures or abstract concepts.

Guidance from the USPTO

From 2014 to 2016 and again in 2019, the U.S. Patent and Trademark Office (USPTO) issued guidance aimed at clarifying patent eligibility criteria. This included illustrative examples of patent claims that were either accepted or rejected, helping practitioners navigate the complex terrain following the Alice decision.

The Rise of AI: A New Challenge

As the influence of AI technologies surged, the implications of the Alice framework became increasingly pertinent. The core functionality of AI—such as generating predictions or responses akin to human thought—raises unique challenges. Courts and the USPTO often regard the ability to execute tasks that a human could perform with basic tools as falling into the realm of abstract ideas.

Consequently, the USPTO has adapted its guidance, incorporating specifics related to AI innovations, which are characterized by their dynamic interaction with existing technological processes.

Understanding Recent Examples for Patent Applicants

For technology startups focused on the AI sector, leveraging recent examples from the USPTO can be critical for securing patent protection. Two notable examples in 2025 illustrate effective strategies:

• Example 47 – Anomaly Detection:

  This example showcases a claim related to software-based AI that includes specific hardware components. By detailing the system architecture and its implementation, applicants stand a better chance of demonstrating non-abstract features. Claims that simply describe functions performed by a computer, without any specifics on improvements or implementations, may be at risk of rejection.

• Example 48 – Speech Separation:

  This claim was noted for its practical application, creating a novel speech signal devoid of unwanted noise. This underscores the importance of illustrating how AI enhances existing processes or applications, thus framing the innovation as a concrete improvement rather than a vague idea.