Artificial Intelligence (“AI”) is revolutionizing the world, akin to past innovations such as electricity and the internet, profoundly impacting global society. AI stands as one of the most transformative, beneficial, and concerning technologies of our time, and while some embrace its potential for improving humanity, others caution against its perceived threats, echoing the same debates attending the emergence of the internet and other potentially disruptive innovations.
The financial world demonstrates a keen interest in AI, with Venture Capital firms continuing to invest in AI companies despite challenging market conditions. The evolving narrative around AI- marked by events like Elon Musk’s lawsuit against OpenAI for abandoning their original mission of being a non-profit, and growing global attention towards managing AI’s power through government oversight- fuels a dynamic landscape fraught with legal challenges.
It is important for board members to identify potential risks to the companies they serve. Use or misuse of Generative AI (“GenAI”) could lead to not only reputational risk, but also financial risk. Consider the risks of not complying with new and emerging government controls as well as the potential for copyright infringement.
Recent developments include:
- The UN’s groundbreaking resolution on AI, garnering agreement from all 193 member countries.
- Nine class-action lawsuitsin the US targeting AI firms, including OpenAI, Meta, and Alphabet.
- The New York Times filing a lawsuit against OpenAI for copyright infringement
- The New York Times lawsuit was headlined by The Verge in a recent article
According to the article, “the stakes are very high…this is a potential extinction-level event for the modern AI industry on the level of what Napster and fellow file sharing sites were facing in the early 2000s. And as we know from history, the copyright rulings from the file sharing age made entire companies disappear, and copyright was changed forever.”
“Today many of those lawsuits are aimed at the companies providing Generative AI, as they are collecting the data into their AI engines.”
Let’s focus on copyright infringement risks which can be unwittingly magnified when GenAI is used. So, just how serious is copyright infringement:
- Copyright infringement is a felony.
- It is possible, though historically shown to be unlikely, that someone could go to jail for copyright infringement.
- Even if the settlement is low, the cost of legal proceedings can be very expensive, and time consuming.
- A lawsuit could negatively impact your brand, and hurt revenues and profits.
There is a difference between copyright infringement and “fair use.” Fair use allows limited portions of the work to be referenced such as in quotes, commentary, criticism, parody, news reporting, and scholarly documents.
However, as a user of generative AI (GenAI), we have no way to tell where the data came from. How is the company you serve using GenAI? I’ve been using ChatGPT for a while, and I’ve yet to ever have it provide the source, or footnotes of where the information came from in its response. We know the data was scraped from somewhere, but knowing from where exactly is almost impossible to know.
GenAI doesn’t typically make things up, though I’ve read up to 27% of responses are hallucination.. And that’s another area companies need to be aware of- GenAI is known to “hallucinate” based on misinterpreting the data, presenting false information, and offering it as fact. Avoiding hallucinations requires humans to review GenAI responses for accuracy and validation.
In either case, copyright infringement or hallucinations can cause companies serious concerns. Unintentionally, it can be too easy to fall prey to copyright infringement, because all the data came from some source, and you have no way of knowing what was copyrighted and what is available for fair use.
What is the Board’s Responsibility?
It can readily be stated that the use of GenAI introduces another, potentially new, risk element to an enterprise making use of the technology. The Board must be aware of the risk of copyright infringement’s potential impact on all companies, without regard to intent. Employing GenAI carries inherent risks, particularly when these platforms face legal action for copyright violations. As a consequence, companies must scrutinize their GenAI usage across departments like R&D, marketing, sales and HR, even considering its use in the development of normal business documents such as agreements, ensuring compliance to avoid costly litigation.
The Board’s responsibility is paramount. Given GenAI’s reliance on data scraped from the Internet and various other sources (recordings, existent documents, fair use materials, etc.), inadvertent copyright infringement is a genuine concern. Establishing robust GenAI Risk Policies such as a GenAI Review Committee, GenAI Governance System, GenAI Policies and Guidelines, and a GenAI Dashboard is crucial. These measures not only mitigate some risks, but also serve to demonstrate proactive efforts to comply with copyright laws.
Lawsuits pose significant financial burden, even for innocent companies. By implementing comprehensive policies and demonstrating proactive compliance measures, companies can reduce risks and potentially defend against allegations of copyright infringement.
Asking questions and seeking clarity in navigating GenAI’s legal landscape is imperative.
I asked ChatGPT 3.5 the following question.
“What's the risk to a company of copyright infringement from the use of generative AI?”
ChatGPT
The risk to a company of copyright infringement from the use of generative AI can be significant, depending on how the AI is being used and the source material it's trained on. Here are some key considerations:
- Use of copyrighted material for training: If a company uses copyrighted material as part of the training data for its generative AI models without proper authorization or licensing, it may be infringing on the copyright holder's rights. This could lead to legal action and potential liability for damages.
- Creation of derivative works: Generative AI models can produce outputs that may be considered derivative works based on copyrighted material. If these outputs are distributed or used commercially without permission, it could constitute copyright infringement.
- Ownership of generated content: There may be questions about the ownership of content generated by AI. In some jurisdictions, the creator of a work is considered the owner of the copyright. However, if the AI is created and controlled by the company, the company may assert ownership over the generated content. Nevertheless, this could still raise legal challenges and disputes, especially if the generated content closely resembles copyrighted material owned by others.
- Fair use considerations: Companies must also consider whether their use of generative AI falls under the doctrine of fair use, which allows for certain limited uses of copyrighted material without permission. Factors such as the purpose and character of the use, the nature of the copyrighted work, the amount and substantiality of the portion used, and the effect of the use on the potential market for the copyrighted work will be considered in determining whether a particular use qualifies as fair use.
To mitigate the risk of copyright infringement when using generative AI, companies should:
- Ensure that they have proper authorization or licensing for any copyrighted material used in training their AI models.
- Implement policies and procedures for evaluating the legality of using generated content, especially if it resembles existing copyrighted material.
- Consult with legal experts to assess the copyright implications of their use of generative AI and to develop strategies for managing and mitigating potential risks.
As a board member, it's essential to ensure that for the companies you serve, that the board and the officers uphold their Duty of Care by actively investigating, asking pertinent questions, ensuring you are well informed, striving to fully comprehend the issues at hand, and remaining alert to potential concerns and problems. It's crucial to establish appropriate oversight mechanisms to mitigate the risks associated with all aspects of AI, including the potential for copyright infringement.
ABOUT PATRICIA WATKINS
Patricia Watkins is an experienced board member, Go-To-Market (GTM) Strategist and Sales Growth Expert. She has held senior leadership roles in Sales, Marketing, Alliances, and Channels, with Fortune companies including HP, Teradata, AT&T, NCR, and a number of start-ups in Silicon Valley. Patricia has led new teams starting at $0 million to existing teams delivering in excess of $800 million in sales.
She is currently an Independent Board Director on 1 public board, 1 private board, and she is on 4 Advisory Boards.
She is the #1 Amazon best-selling author of two books, Driving More Sales, 12 Essential Elements, and Land and EXPAND, 6 Simple Strategies to Grow Your Top and Bottom Line.
She graduated with a BBA from The University of Texas, and an MBA from Santa Clara University, both with honors.
Disclaimer: The views and opinions expressed in this blog are solely those of the authors providing them and do not necessarily reflect the views or positions of the Private Directors Association, its members, affiliates, or employees.