OpenAI’s Shift from Trade Secrets to Patent Protection

OpenAI, a leading player in the artificial intelligence (AI) sector, has historically relied on trade secrets to protect its technological advancements. However, recent developments indicate a major strategic shift toward patent protection. We explore the implications of this shift and the potential impact on business, from small startups to large corporations.

From non-profit to commercial

In 2022, the OpenAI became a household name with the release of ChatGPT, a large language model (LLM) capable of generating human-like text. However, OpenAI began as a non-profit organization in 2015 with the goal of developing safe and beneficial AI for humanity. Since then, OpenAI has taken steps towards a more commercial model by creating its profit-based subsidiary, OpenAI Global, LLC. Famously, Microsoft has invested approximately $13 billion into OpenAI, and there are rumours that OpenAI is planning a full transition to a profit-based model. A switch from trade secrets to a patent heavy IP portfolio would certainly fit well into this profit-based shift. Watch our webinar to find out more about trade secrets here.

The Trade Secret Approach: What are the pros and cons?

Trade secrets have traditionally been OpenAI’s preferred method for safeguarding its IP. For example, while ChatGPT has arguably democratized the use of AI by making it accessible in natural language, little is know about the technological details that make it so successful.

The advantages of trade secrets include:

1. Indefinite Protection: Trade secrets can protect technologies indefinitely, provided they remain confidential.

2. No Disclosure Requirement: Companies can keep their innovations hidden from competitors, avoiding any mandatory public disclosure that patents require.

3. Cost-Effective: Maintaining a trade secret is less expensive than obtaining and enforcing a patent, which can involve substantial upfront costs and ongoing maintenance fees.

However, a reliance on trade secrets comes with significant risks:

1.Reverse Engineering: Competitors can legally reverse-engineer a product and recreate the technology, thus circumventing trade secret protections. If someone works out how to copy important aspects of your innovation, it is difficult to stop them using it.

2.Employee Turnover Risks: Employees with access to sensitive information might inadvertently or deliberately disclose trade secrets, leading to a potential loss of IP protection. In areas with limited talent pools, such as in machine learning, this can be especially dangerous.

3. Enforcement Challenges: Proving misappropriation of trade secrets in court is difficult, and the legal remedies may be less effective compared to patent infringement cases.

Why the Strategic Shift to Patent Protection?

OpenAI’s decision to pivot towards patent protection could reflect the growing competitive pressures in the AI industry, and the need for more enforceable IP rights. Meta’s LlaMA3, Google’s Gemini 1.5 and Anthropic’s Claude 3 are just some examples of famous LLMs in the same competitive market.

Patents offer several strategic advantages:

1. Exclusive Rights: Patents grant OpenAI exclusive rights to its innovations for a fixed period, typically 20 years, preventing competitors from using, selling, or licensing similar technologies.

2. Stronger Legal Enforcement: Patent rights are more straightforward to enforce in court, providing a more robust defense against infringement.

3. Revenue Generation: Patents can be monetized through licensing or sales, creating new revenue streams and strategic opportunities, particularly in cross-licensing agreements.

4. Valuation: An EU Study showed startups with patents increase their chances of securing funding by over 600%. Read more about the impact of IP on startup valuations here.

As OpenAI’s technologies are increasingly integrated across industries, a robust IP protection strategy is essential to maintaining a competitive edge.

Patentability of AI: Another Legal Milestone

The recent patents secured by OpenAI underscore that AI technologies are indeed patentable, provided they meet the requisite criteria of novelty, inventive step, and industrial applicability. Each of OpenAI’s patents demonstrates that AI systems, especially those involving complex algorithms, machine learning models, and innovative applications of these models, can be protected under patent law.

These patents, filed in the United States throughout 2023, cover AI-driven technologies from text and image generation and editing to code generation to integration with external systems. These patents confirm that AI, when applied in novel and non-obvious ways, meets the criteria for patentability, helping cement a legal precedent that could help influence future AI-related patents. If you are especially interested in OpenAI’s patented technologies, contact me for a more in-depth review!

Rapid Patenting and Anticipated European Applications

One of the most fascinating aspects of OpenAI’s new approach is its focus on rapid patenting. Traditionally, it can take 3 to 5 years to get a patent granted, but OpenAI has chosen to pursue an accelerated process. This often involves filing direct national applications without claiming priority, which can result in a grant in less than a year.

All of the published applications from OpenAI are U.S. cases, and nine of these have already been granted. Some of these were granted as quickly as nine to ten months after filing. This strategy could suggest that OpenAI intends to enforce its patents or use them as negotiating chips in licensing discussions.

Given that these patents were filed in the U.S. throughout 2023, it is reasonable to expect that corresponding international (PCT) or European Patent Office (EPO) applications are either in progress or will be filed soon. Patent applications typically publish 18 months after the filing date, so we anticipate the international equivalents will be published later in 2024 and into 2025.

If you want to explore the possibility of expedited processing for your patent portfolio, contact us here.

The likelihood of OpenAI’s patents being granted by the EPO is difficult to estimate, as they may face stricter scrutiny and require modifications. The EPO is known for its rigorous approach to AI-related patents, particularly emphasizing the inventive step and technical contribution of the invention beyond abstract algorithms. While OpenAI’s patents, which involve complex machine learning models and innovative applications, have a strong foundation for patentability, the EPO may require adjustments to claims to clarify their technical effects. Therefore, while these patents are likely to be granted, it is probable that some claims will need to be narrowed or redefined to meet the EPO’s stringent standards.

For organisations and enterprises in Europe, including in Norway and the UK, these forthcoming patents will have significant implications. OpenAI has the market reach and finances to potentially seek protection across many countries involved in AI, meaning these patents will likely extend into Norway as well. Entities operating in AI should be prepared to navigate this emerging landscape, considering how these patents may affect their operations, freedom to operate, and potential for collaboration or licensing.

Impact on the AI Market and the Research Economy

If patents for these technologies are granted across Europe in the coming years, it will potentially have a ripple effect across the AI market and the broader research economy.

Such a strong patent portfolio could:

1. Drive Innovation: By securing patents, OpenAI encourages further innovation as competitors and collaborators alike strive to develop new, non-infringing technologies.

2. Influence Market Dynamics: OpenAI’s patents will set new benchmarks in the AI market, potentially raising barriers to entry for smaller competitors or fostering partnerships through licensing agreements.

3. Shape Research Directions: The patents may steer academic and corporate research towards unexplored areas, particularly those not covered by OpenAI’s IP. This could lead to increased specialization and differentiation in AI research.

For European businesses, and across the world, these developments present both challenges and opportunities. On one hand, the patents may limit the use of certain AI technologies, necessitating careful freedom-to-operate analyses before launching new products or services. On the other hand, these patents create opportunities for licensing agreements, joint ventures, or R&D collaborations with OpenAI, enabling organisations to leverage cutting-edge AI technologies without developing them independently.

It is also a great opportunity to explore previously hidden technology as the patents become public – an important cornerstone of the patent system across the world is the ability to understand and learn from new inventions!

What are the Strategic Considerations for Businesses?

Businesses and organisations in the AI sector, both big and small, should consider the following strategies in response to OpenAI’s expanding patent portfolio:

1. Patent Landscaping and Freedom to Operate: Conduct thorough patent landscaping to understand the scope of OpenAI’s patents and identify areas of potential infringement risk or opportunities for innovation.

2. IP Portfolio Development: Focus on building a strong IP portfolio by filing patents for any new and inventive AI technologies, ensuring protection and increasing bargaining power in negotiations.

3. Collaboration and Licensing: Explore opportunities to license patented technologies from OpenAI or collaborate on joint ventures that leverage both parties’ strengths.

4. Research and Development: Invest in R&D to innovate around existing patents, potentially creating new technologies that complement or improve upon OpenAI’s offerings.

Conclusion

OpenAI’s move from relying on trade secrets to building a comprehensive patent portfolio marks a significant evolution in its IP strategy. This shift not only underscores the patentability of AI but also sets the stage for a more competitive and innovative AI market. For organisations and enterprises in Europe, including those in Norway, this presents both challenges and opportunities. By staying informed and strategically adapting to this evolving landscape, businesses can position themselves to thrive in the rapidly expanding AI economy.