accessiBe Ltd v. AudioEye Inc

1. Case Identification

• Case #: IPR2022-00226
• Patent #: 10,866,691
• Filed: December 15, 2021
• Petitioner(s): accessiBe Ltd.
• Patent Owner(s): AudioEye, Inc.
• Challenged Claims: 1-30

2. Patent Overview

• Title: Modular Systems and Methods for Selectively Enabling Cloud-Based Assistive Technologies
• Brief Description: The ’691 patent discloses systems and methods for automatically remediating website accessibility issues. The process involves a remote computer (client) sending a remediation request to a server, which in turn provides remediation code (e.g., a script) that is executed by the client's browser to fix compliance issues by modifying the webpage's Document Object Model (DOM) or HTML.

3. Grounds for Unpatentability

Ground 1: Obviousness over Springer and Hendry - Claims 1-9, 11-13, 16-24, and 26-28 are obvious over Springer in view of Hendry.

• Prior Art Relied Upon: Springer (Application # 2004/0148568) and Hendry (Application # 2013/0104029).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Springer taught a system for automatically detecting and fixing web accessibility compliance issues. This system used "checker" tools to find violations (e.g., images lacking 'alt' attributes) and corresponding "fixer" tools to automatically add the required attributes to the webpage's HTML or DOM. Petitioner asserted that Hendry taught a distributed, client-server architecture for improving document accessibility, where a server provides remediation code (e.g., JavaScript) to a client machine, which then executes the code in its browser to apply accessibility features. The combination allegedly disclosed the claimed method of a remote computer requesting and receiving remediation code from a server to fix accessibility issues.
  • Motivation to Combine: A POSITA would combine these references to implement Springer's accessibility remediation logic within Hendry's more efficient and modern client-server framework. This combination would yield predictable benefits such as improved scalability, performance, resiliency, and the ability to provide up-to-date remediation centrally from the server.
  • Expectation of Success: A POSITA would have expected success in combining the references because both Springer and Hendry operated on the same fundamental principle of modifying a webpage's HTML/DOM to remedy accessibility problems, making their teachings compatible.

Ground 2: Obviousness over Springer, Hendry, and Nguyen - Claims 10, 14, 15, 25, 29, and 30 are obvious over Springer in view of Hendry and Nguyen.

• Prior Art Relied Upon: Springer (Application # 2004/0148568), Hendry (Application # 2013/0104029), and Nguyen (a 2008 journal article titled “Learning to Extract Form Labels”).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground built upon the combination of Springer and Hendry by adding the teachings of Nguyen. Petitioner contended that Nguyen disclosed using artificial intelligence (AI) and machine learning techniques to improve the process of identifying labels for untagged form fields by analyzing contextual cues within a webpage. This teaching was argued to render obvious the claims reciting the use of AI, machine learning, or the analysis of contextual cues to generate remediation code.
  • Motivation to Combine: A POSITA would have been motivated to incorporate Nguyen’s AI/ML techniques into the Springer/Hendry system to improve its accuracy. Nguyen’s methods were presented as a known improvement over the older, heuristics-based approaches for identifying non-textual elements, offering a more robust way to generate accurate descriptions and remediation code, particularly for complex elements like form fields.
  • Expectation of Success: Success was expected because AI and remote computing techniques were well-known at the time of the invention. Integrating Nguyen's AI-based label extraction into the server-side processing of the Springer/Hendry framework was a predictable application of known technologies to improve a known system.

Ground 3: Obviousness over Lehota and Hendry - Claims 1-9, 11-13, and 16-28 are obvious over Lehota in view of Hendry.

• Prior Art Relied Upon: Lehota (Application # 2010/0205523) and Hendry (Application # 2013/0104029).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground presented Lehota as an alternative base reference to Springer. Petitioner argued Lehota disclosed a system that used an "accessibility widget" to scan a website's DOM, find inaccessible elements (e.g., images missing ALT tags), and modify the DOM to make the website more accessible. The core argument mirrored Ground 1, asserting that the combination of Lehota’s widget-based remediation logic with Hendry's client-server architecture taught the claimed invention.
  • Motivation to Combine: The motivation was identical to that in Ground 1: a POSITA would implement Lehota's remediation system in Hendry's distributed computing environment to gain the well-understood benefits of a client-server architecture, such as improved performance and centralized management of remediation logic.
  • Expectation of Success: A POSITA would have expected the combination to work, as both Lehota and Hendry taught remediation techniques that relied on modifying the DOM, ensuring their compatibility.
• Additional Grounds: Petitioner asserted an additional obviousness challenge (Ground 4) against claims 10, 14, 15, 25, 29, and 30 based on Lehota in view of Hendry and Nguyen, which mirrored the logic of Ground 2 by adding Nguyen’s AI/ML teachings to the Lehota/Hendry combination.

4. Key Claim Construction Positions

• Petitioner noted that the district court in a parallel litigation construed the terms “descriptive alt text”/“alt text description” to mean “descriptive alternative text”/“alternative text description.” Petitioner stated it applied this construction in its analysis while reserving the right to respond to any alternative constructions proposed by the Patent Owner.

5. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under Fintiv would be inappropriate. The petition asserted that the trial date in the parallel district court proceeding was uncertain and likely to slip, minimizing overlap with the IPR timeline. Petitioner further argued that investment in the court proceeding had been minimal, as no prior art-based invalidity arguments had been presented to the court. Finally, Petitioner contended that the strong merits of its unpatentability challenges weighed heavily in favor of institution.

6. Relief Requested

• Petitioner requested institution of an inter partes review and cancellation of claims 1-30 of the ’691 patent as unpatentable under 35 U.S.C. §103.