“Generative AI” or GenAI is quickly becoming a household term, but how many organizations understand its benefits and risks?
GenAI is, at its most basic, artificial intelligence that generates content according to whatever instructions the user feeds the software. These instructions (or prompts) direct GenAI to create text, audio, images, video or other material. Some examples include “create an image of a lion with wings in a garden,” or “write lyrics similar to one of Taylor Swift’s songs about how great dogs are.”
While some GenAI produces only one kind of output – such as images or music – all GenAI are trained based on existing content or data, the source of which is often unknown to the user. In fact, many GenAI models utilize the user’s input to further train and improve itself. This is both beneficial and risky. To establish best technical practices several security agencies have issued joint guidelines. In this blog we consider some of the pros and cons of GenAI from a legal risk management point of view.
Benefits
The applications of GenAI are endless and have the potential to revolutionize business processes. GenAI is efficient. Quicker than humans, GenAI can take plain facts and create exciting marketing copy to launch a product. Other GenAI can aid in research and design in, say, testing software. GenAI can even be used for learning, such as guiding an individual through a programming problem and providing an explanation with examples.
Risks
A key risk of using GenAI is the potential loss of confidentiality or proprietary rights in information inputted into the GenAI. As noted, some forms of GenAI learn from user prompts, which in turn may be used in outputs for future users. As a result, prompts that contain confidential, proprietary or sensitive information may be used repeatedly by the GenAI to fulfill other user requests. GenAI may also store this information in repositories, perhaps overseas or with third parties, which are beyond your control. This can raise serious concerns over loss of confidentiality or proprietary rights of your information.
Another significant area of risk is questions around the ownership of the output created by the GenAI. By virtue of the generative nature of GenAI, output may not be unique and in some cases can be the same or similar to output generated by other users or the very data that it has been trained on. This raises serious concerns around the ownership of intellectual property (IP) rights in the output and could result in copyright infringement claims.
In Canada, the law in this respect remains under development so it is currently unclear whether a person can be the author of AI-generated work after merely entering prompts. Globally, questions have also been raised asking whether GenAI can be considered an inventor under patent law.
Other concerns include inaccurate or incomplete data, misinformation or outright disinformation, and reflecting social bias (such as racial or gender bias). There may also be vulnerabilities relating to cybersecurity and intentional attacks on the GenAI system, such as “poisoning” (corrupting) a system with malware.
Essential questions to ask
What policies should govern GenAI use at your organization? How should you address AI and preserve legal rights in contracts? How should sellers and purchasers of a business clarify and ascertain IP ownership of materials created using GenAI? Does your organization have a policy or strategy for addressing GenAI use? Are your employees properly trained and informed on the risks and benefits of GenAI? Is your organization following the most current AI laws, regulations and guidelines?
The answers are complex, yet crucial for your organization. Our innovation, data and technology team at MLT Aikins is happy to guide you through the essentials of GenAI at the third annual MLT Aikins IP and Technology Conference on June 4, 2024 in-person at the Radisson Hotel Winnipeg Downtown or virtually.
Note: This article is of a general nature only and is not exhaustive of all possible legal rights or remedies. In addition, laws may change over time and should be interpreted only in the context of particular circumstances such that these materials are not intended to be relied upon or taken as legal advice or opinion. Readers should consult a legal professional for specific advice in any particular situation.