Sema AI Working Paper 02: Assessing IP risks of coders using GenAI

Executive Summary and Introduction

Sema AI Working Paper 01 found a significant ROI (up to 41X over two years) for coders using GenAI tools such as GitHub CoPilot and ChatGPT to develop software faster.

These benefits are only worth the investment if risks from coders using GenAI can be mitigated. One of the biggest potential risks is receiving Intellectual Property protection for the generated code.

Sema’s research indicates that the IP can indeed be protected, and the business risk is low when GenAI is used corrected.

1. Risk of legal impact from creators of GenAI training data to users of GenAI tools is low.
2. GenAI in the Software Development Lifecycle (SDLC) does not present a meaningful risk to receiving trade secret protection when the proper tools are used correctly.
3. GenAI in the SDLC does not present a meaningful risk to receiving US patent protection.
4. GenAI in the SDLC presents some medium term risk for receiving copyright protection. Companies that plan to seek copyright protection should work proactively with Counsel to prepare and should plan to disclose GenAI usage.
   

Considering the above, as well as the ability to mitigate the additional significant risks from coders using GenAI [Data Leakage, Security Risks, Quality Risks—covered in Working Paper 04] there are two paths to consider:

1. If an organization determines that the benefits from GenAI developer tools outweigh the risk, organizations should:some text
   1. Adopt Enterprise-grade GenAI tools.
   2. Ban the use of other tools (“bring your own LLM license to work”).
   3. Proactively work with Counsel to prepare both for potential disclosure requirements (i.e. acknowledging use of GenAI in the application, if needed) and for protecting the quality of the application(s).
2. If an organization determines that the benefits from GenAI developer tools do not outweigh the risk, then organizations should both ban the use of such tools and monitor the code to detect then prevent usage.
   

Sema strongly recommends Option 1.

Note this is business judgment advice, not legal advice, and organizations should make a usage determination with their Counsel.

‍

‍

1. Risk of legal impact from creators of GenAI training data to users of GenAI tools is low.

Some GenAI tools have been trained on data whose creators are asserting inappropriate use. Examples of this include lawsuits from the New York Times and members of the Open Source community.

Sema assesses the risk of legal action against creators of GenAI tools to be Critical, i.e. those risks are already realized or could be imminent.

Nonetheless, Sema assesses the risk of legal action against users of GenAI coding tools to be low, i.e. those risks are unlikely to be realized.

Sema comes to this conclusion for the following reasons:

1. The plaintiffs estimate that “that ‘about 1% of the time’ Copilot generated snippets of code similar to the publicly available code from which it learned.” This low percentage of directly copied code lowers the likelihood of successful action against coding tool users.
2. At least one GenAI coding tool, GitHub Copilot Enterprise Cloud, offer settings that further reduce the risk of reuse by prohibiting snippets of reused code.
3. Major GenAI coding providers have offered protection against legal action against users: GitHub Copilot, OpenAI, and CodeWhisperer.
4. Historical legal action dynamics, including but not limited to the rarity of threatened legal action over Open Source license misuse, also suggest a low likelihood of legal action in this situation.
   

If an organization proceeds with GenAI coding tools, the following steps are recommended to minimize the risk of legal action from creators:

1. Pick an Enterprise-grade tool and license that either includes indemnification protections or was not trained on proprietary code.
2. If using GitHub Copilot, disable snippets of reused code.
3. Be a “good corporate citizen” with respect to the Open Source community, such as by contributing developer time or financial resources to Open Source packages that the company uses directly.

‍

‍

2. GenAI in the Software Development Lifecycle (SDLC) does not present a meaningful risk to receiving trade secret protection when the proper tools are used correctly.

Companies receive trade secret protection by keeping their information inaccessible to the public.

Unlike copyright, patent, and trademark law-each of which require the IP-owner to disclose their IP to place the public on notice of the owner's rights-trade secret law requires the IP to be kept, well, "secret." The Defend Trade Secrets Act (the federal trade secret law) defines a "trade secret" to include "all forms and types of ... business. information" so long as "the owner thereof has taken reasonable measures to keep such information secret" and "the information derives independent economic value, actual or potential, from not being generally known to, and not being readily ascertainable through proper means ... . " 18 U.S.C. $ 1839(3). https://www.huschblackwell.com/newsandinsights/a-legal-primer-on-artificial-intelligence-and-intellectual-property

Code created by GenAI tools may not get trade secret protection:

If information claimed to be trade secret was originally created in whole or in part by Gen-Al, it is significantly less likely to be considered a "trade secret." Husch Blackwell, A Legal Primer on Artificial Intelligence and Intellectual Property, March 11, 2024

Companies should ensure that materials submitted to GenAI tools are submitted secretly, such as by using private GenAI tools.

Moreover, if trade secret information is publicly disclosed when used to train Gen-Al, the information will likely be deemed to no longer meet the trade secret definition. Extreme care should be taken, therefore, to protect against the accidental disclosure of confidential information when using Gen-AI to avoid destroying trade secrets Husch Blackwell, A Legal Primer on Artificial Intelligence and Intellectual Property, March 11, 2024 https://www.huschblackwell.com/newsandinsights/a-legal-primer-on-artificial-intelligence-and-intellectual-property

‍

‍

3. GenAI in the SDLC does not present a meaningful risk to receiving US patent protection.

Organizations seeking US patent protection do not appear to have an affirmative requirement to disclose their use of GenAI if the GenAI usage is not a considered barrier to receiving patent protection:

Because improper inventorship is a ground of rejection under 35 U.S.C. 101 and 115,68 parties identified in 37 CFR 1.56(c), 1.555(a), and 42.11(a) have a duty to disclose to the USPTO information that raises a prima facie case of unpatentability due to improper inventorship or that refutes, or is inconsistent with, a position an applicant takes in opposing an inventorship rejection or asserting inventorship. For example, in applications for AI-assisted inventions, this information could include evidence that demonstrates a named inventor did not significantly contribute to the invention because the person’s purported contribution(s) was made by an AI system. U.S. Patent and Trademark Office, Inventorship guidance for AI-assisted inventions February 12, 2024 https://www.govinfo.gov/content/pkg/FR-2024-02-13/pdf/2024-02623.pdf

At this time, to meet their duty of disclosure, applicants rarely need to submit information regarding inventorship. The USPTO does not believe this inventorship guidance will have a major impact on applicants’ disclosure requirements. However, special care should be taken by those individuals subject to this duty to ensure all material information is submitted to the USPTO to avoid any potential negative consequences. U.S. Patent and Trademark Office, Inventorship guidance for AI-assisted inventions February 12, 2024

Organizations that use GenAI coding tools can still receive US patent protection:

Additionally, there are no other sections of the Patent Act that support a position that inventions that are created by natural person(s) using specific tools, including AI systems, result in improper inventorship or are otherwise unpatentable. The statutes only require the naming of the natural persons who invented or discovered the claimed invention, irrespective of the contributions provided by an AI system or any other advanced system. U.S. Patent and Trademark Office, Inventorship guidance for AI-assisted inventions February 12, 2024

The use of an AI system by a natural person(s) does not preclude a natural person(s) from qualifying as an inventor (or joint inventors) if the natural person(s) significantly contributed to the claimed invention, as explained in section IV of this notice. U.S. Patent and Trademark Office, Inventorship guidance for AI-assisted inventions February 12, 2024

‍

‍

4. GenAI in the SDLC presents some medium-term risk for receiving US copyright protection. Companies that plan to seek copyright protection should work proactively with Counsel to prepare and should plan to disclose GenAI usage.

Organizations seeking US copyright protection appear to have an affirmative requirement to disclose their use of GenAI:

Consistent with the policies described above, applicants have a duty to disclose the inclusion of AI-generated content in a work submitted for registration and to provide a brief explanation of the human author’s contributions to the work. US Copyright Registration Guidance, March 16, 2023 https://www.govinfo.gov/content/pkg/FR-2023-03-16/pdf/2023-05321.pdf

Applicants should not list an AI technology or the company that provided it as an author or co-author simply because they used it when creating their work. US Copyright Registration Guidance, March 16, 2023

Allen also asserted that “[r]equiring creators to list each tool and the proportion of the work created with the tool would have a burdensome effect if enforced uniformly,” but the Board rejected this argument too, noting that the disclosures don’t have to be detailed or specific, but only require a brief statement (e.g., “the text was generated by artificial intelligence”). Perkins Coie, Human Authorship Requirement Continues To Pose Difficulties for AI-Generated Works, February 29, 2024 https://www.perkinscoie.com/en/news-insights/human-authorship-requirement-continues-to-pose-difficulties-for-ai-generated-works.html

Applicants who fail to update the public record after obtaining a registration for material generated by AI risk losing the benefits of the registration. If the Office becomes aware that information essential to its evaluation of registrability ‘‘has been omitted entirely from the application or is questionable,’’ it may take steps to cancel the registration.44 Separately, a court may disregard a registration in an infringement action pursuant to section 411(b) of the Copyright Act if it concludes that the applicant knowingly provided the Office with inaccurate information, and the accurate information would have resulted in the refusal of the registration. US Copyright Registration Guidance, March 16, 2023

Under certain circumstances, use of GenAI coding tools may lead to code that cannot receive copyright protection:

The Office’s current official policy, published on March 10, 2023, is that it will register a work only if the work’s traditional elements of authorship were authored by a human and not by a machine.² The Office distinguishes between works autonomously generated by AI, which are not protectable by copyright, and works created with the assistance of AI, for which a case-by-case analysis is necessary to determine whether the expressive elements are the product of a human or of a machine. Reed Smith, Legal issues of AI in the entertainment and media sector part 1, February 2024 https://www.reedsmith.com/en/perspectives/ai-in-entertainment-and-media/2024/02/ip-copyright

In the case of works containing AI-generated material, the Office will consider whether the AI contributions are the result of ‘‘mechanical reproduction’’ or instead of an author’s ‘‘own original mental conception, to which [the author] gave visible form.’’ 24 The answer will depend on the circumstances, particularly how the AI tool operates and how it was used to create the final work.25 This is necessarily a case-by- case inquiry. US Copyright Registration Guidance, March 16, 2023

‍

‍

For More Information

Relationship to other Working Papers:

• Working Paper 01- High ROI AI Activities. Explains the calculation behind the ROI for coders using GenAI.
• Working Paper 03- Comparison of tiers of GitHub Copilot GenAI Coding Tools. A critical method to reduce risks from coders using GenAI tools is to manage access to the appropriate tools. Working Paper 03 compares GitHub Copilot’s offerings and recommends CoPilot Enterprise.
• Working Paper 04- Benefit / Risk Assessment of Using GenAI Coding Tools. In total, considering the benefits and the mitigatable nature of the risks of GenAI tooling, it is strongly recommended to adopt GenAI coding tools.
  

Contact ai@semasoftware.com.

‍

‍

Disclosure

Sema publications should not be construed as legal advice on any specific facts or circumstances. The contents are intended for general information purposes only. To request reprint permission for any of our publications, please use our “Contact Us” form.

The availability of this publication is not intended to create, and receipt of it does not constitute, an attorney-client relationship. The views set forth herein are the personal views of the authors and do not necessarily reflect those of the Firm.