Content-Generating AI
Content-generating AI can potentially disrupt a wide range of knowledge and intelligence-based industries. OpenAI has captured the public’s imagination, but its lack of a defensive patent strategy, coupled with a projected revenue of over $1 billion in the near term, exposes it to material risks of inbound assertion from innovators and invention owners in this space.
Microsoft announced in January 2023 the third phase of a long-term alliance with OpenAI to accelerate AI breakthroughs. This investment, rumoured to be around $10 billion, follows previous investments made in 2019 and 2021 and extends the companies’ collaboration across AI super-computing and research.
So, who are OpenAI, and what are their aims?
OpenAI describes itself as an AI research and deployment company with a mission to ensure that artificial general intelligence benefits all of humanity. The company was founded by tech investors including Peter Thiel, Elon Musk, Reid Hoffman and Sam Altman in 2015, with headquarters in San Francisco. OpenAI recently went viral following the November 2022 launch of a free preview of its AI chatbot, ChatGPT.
What is ChatGPT?
GPT stands for Generative Pre-trained Transformer. It is based on a language model called GPT-3, which is trained on large open datasets and aimed at generating natural language answers to questions, but with other functions including translation and improvised text generation. With OpenAI’s projected revenues forecast to exceed $1bn by 2024, ChatGPT-3 has the potential to open up a whole new market around content-creating AI.
The chatbot has launched a completely new way to interact with information and generate complex answers to a myriad of questions.
OpenAI – and open source
Interestingly, OpenAI does not appear to have any published patents, pending or granted. In a blog post titled ‘Introducing OpenAI’ published in 2015, the company founders shared their aims, including
- ‘As a non-profit, we aim to build value for everyone rather than shareholders. Researchers will be strongly encouraged to publish their work, whether as papers, blog posts, or code, and our patents (if any) will be shared with the world.
- ‘We’ll freely collaborate with others across many institutions and expect to work with companies to research and deploy new technologies.’
This open-sourced approach may work for a non-profit with idealistic goals. However, it now presents risks to OpenAI’s commercial strength, given its shift in 2019 towards a for-profit model. Patents are a great way to demonstrate priority and to protect against assertions from both non-practising entities (NPEs) and technology competitors.
Patents are a key part of a technology-rich IP strategy. OpenAI is likely to be an outlier in high-tech due to its lack of appetite for patents. Most organisations, particularly for-profits, are proactive in protecting future revenues against license assertions and take steps to develop an intellectual property strategy to mitigate that risk.
About the Content Generating AI report
Despite the zero patents identified for OpenAI, the report looks at published patents across Content Generative AI to identify key innovators and invention owners. These organizations represent potential assertion threats to OpenAI, given its high projected revenue, investment from Microsoft and lack of a defensive patent strategy, but many are also likely to be key participants in what appears today a relatively nascent but high-growth, high-tech disruptive market opportunity.
Additionally, to investigate OpenAI’s patenting strategy, the authors asked ChatGPT, to walk Cipher through it. We think that it has summarised the risks which it faces quite well.
Key messages from our report
OpenAI has pursued an open-source strategy in developing its technologies and does not appear to have any published patents at all
This is consistent with its open-source strategy but exposes the company to material risk of inbound assertion associated with patent infringement from patent owners in Content-Generative AI as its revenue grows.
OpenAI has an estimated patent licensing risk of $80 million
This will only grow if they exceed their $1bn revenue prediction.
IBM has the largest portfolio overall, with over 120 families published
But with an average PVIX score of just below 50, their portfolio appears relatively weak compared to their competitors.
Microsoft likely has the strongest generative AI portfolio
This is if we consider their number of grants (66 patents) and their average PVIX, 59. Salesforce and Intel have the highest average PVIX, 67 and 64, respectively, but have far fewer grants (11 and 5 patents).
Microsoft has the strongest generative AI portfolio
This if we consider their number of grants (66 patents) and their average Portfolio Value Index (PVIX), 59. Salesforce and Intel have the highest average PVIX, 67 and 64, respectively, but fewer grants (11 and five patents).
OpenAI has a strong partnership with Microsoft but faces risk from other leading companies in the field, such as Google
With an expected explosion of generative AI offerings in the near future, not all will succeed, which could leave some of these companies looking to monetize their portfolios.
There are currently few relevant patents held by NPEs, indicating the risk is currently quite low
However, as generative AI becomes more profitable, a large pool of small companies, universities, and individual owners may begin selling their patents to NPEs, so there is potential for this risk to increase.
There has been an explosion in patents published in the last five years
This helps explain why the current technology leaders are in their positions. Some portfolios have gone from near-zero to industry-leading, suggesting it is not too late for OpenAI to change strategy.
Download the Content-generating AI report
Sections in this report:
- What can ChatGPT tell us? – Our questions answered by the OpenAI chatbot
- Top Innovators in Generative AI using classified patent data as the primary indicator
- NPE Risk: Low but with the potential to grow
- Patent Filing Trends
- Estimating patent risk exposure
- Patent Classification Methodology