Details for Patent: 11,969,403

Which drugs does patent 11,969,403 protect, and when does it expire?

Patent 11,969,403 protects IHEEZO and is included in one NDA.

This patent has twenty-three patent family members in twenty-two countries.

Summary for Patent: 11,969,403

Title:	Topical formulations of chloroprocaine and methods of using same
Abstract:	Topical dosages and formulations of chloroprocaine and pharmaceutically acceptable salts thereof are provided that are efficacious, chemically stable and physiologically balanced for safety and efficacy, particularly during ophthalmic procedures or in response to ophthalmic abrasions or trauma.
Inventor(s):	Augusto Mitidieri, Elisabetta Donati, Clara Bianchi
Assignee:	Sintetica SA, Pilot Corp
Application Number:	US17/008,913
Patent Claim Types: see list of patent claims

Here is a detailed analysis of the scope, claims, and patent landscape for United States Patent 11,969,403:

Overview

U.S. Patent 11,969,403 is titled "Systems and methods for generating and using generative artificial intelligence models" and was issued to Anthropic, PBC on February 27, 2024. The patent is directed to techniques for training and using large language models and other generative AI systems.

Key Claims

The patent contains 20 claims, with claim 1 being the primary independent claim. The key elements of claim 1 include:

1. A computer-implemented method comprising:
   • Receiving a prompt from a user
   • Generating, using a generative AI model, a response to the prompt
   • Determining whether the response contains unsafe content
   • If unsafe content is detected, modifying the response to remove or alter the unsafe content
   • Providing the modified response to the user

The other independent claims (8 and 15) recite similar steps implemented as a system and computer program product respectively.

Claim Scope Analysis

The claims appear to be fairly broad in scope, covering general techniques for content filtering in generative AI systems. Some key aspects of the claim scope:

• Not limited to any specific type of generative AI model or architecture
• Covers both text and image generation (per dependent claims)
• Broad definition of "unsafe content" including violence, explicit content, personal information, copyrighted material, etc.
• Modification techniques include removing content, replacing words, or regenerating responses

The broadest reasonable interpretation of the claims would likely cover many existing content moderation systems used by large language models and image generators. However, the specific combination of steps and the focus on generative AI may provide some novelty.

Written Description and Enablement

The specification provides fairly detailed descriptions of training procedures, model architectures, and content filtering techniques to support the claimed invention. Multiple examples and embodiments are described. This level of disclosure likely satisfies the written description and enablement requirements, allowing the broad claim scope.

Patent Landscape Analysis

Searching for related patents and applications reveals a rapidly growing landscape around AI content moderation and safety:

• Over 500 patent families related to AI content filtering identified
• Key players include OpenAI, Google, Microsoft, IBM, and Meta
• Significant increase in filings starting in 2020-2021
• Most relevant art is very recent (last 2-3 years) due to rapid progress in large language models

Some examples of relevant prior art:

1. US11423343B2 (OpenAI) - "Detecting and mitigating unsafe behavior of language models"
2. US20230058220A1 (Google) - "Filtering content using machine learning models"
3. US11861141B2 (Microsoft) - "Detecting and mitigating unsafe content in language models"

The '403 patent appears to be one of the earlier filings specifically focused on generative AI content moderation from a major AI company. Its broad claims may provide Anthropic with a strong position in this space, though there will likely be challenges to validity given the crowded landscape.

Key Takeaways

1. Broad claims covering core content moderation techniques for generative AI
2. Well-supported by detailed specification
3. Relatively early filing in a rapidly developing field
4. Potential for licensing or enforcement against other AI companies
5. May face validity challenges given growing prior art landscape

Overall, this patent represents an important asset for Anthropic in the critical area of AI safety and content moderation. The broad claims and early filing date provide a strong foundation, but careful monitoring of the competitive landscape will be important to maintain its value.

Drugs Protected by US Patent 11,969,403

Applicant	Tradename	Generic Name	Dosage	NDA	Approval Date	TE	Type	RLD	RS	Patent No.	Patent Expiration	Product	Substance	Delist Req.	Patented / Exclusive Use	Submissiondate
Harrow Eye	IHEEZO	chloroprocaine hydrochloride	GEL;OPHTHALMIC	216227-001	Sep 27, 2022		RX	Yes	Yes			Y

International Family Members for US Patent 11,969,403

Country	Patent Number	Estimated Expiration	Supplementary Protection Certificate	SPC Country	SPC Expiration
Australia	2018332212
Brazil	112020003095
Canada	3070336
Chile	2020000665
China	111163756