The risks of using AI to create designs
The rise of artificial intelligence (AI) has led to an increased interest in its application to various industries, including the world of graphic and product design.
Whilst AI has many benefits, including automating repetitive tasks, improving design accuracy and speed, as well as generating new ideas and styles, this article from our partner, MFL, comes with a cautionary note.
As with any new technology of this kind, there are potential insurance, intellectual property, and risk management issues that need to be considered. Particularly when the application of that technology is so new, that we are in unchartered territory with little to no legal framework that has been provided by the courts, to decide on issues that arise from its use.
One of the first issues that may need to be considered is the contractual position in respect of the use of AI and there are two aspects to this.
The first relates to the appointment by the Client. They have appointed the designer to provide the services, but if those services are provided by an AI, has the designer ‘really’ fulfilled their contractual obligations?
The second aspect relates to the terms agreed upon between the designer and the providers of the AI service. There are a number of areas that need to be considered; for example, with the benefit of legal advice, you can seek advice on the positions in respect of both liability and copyright.
One of the major concerns of using AI in design is the potential for errors or malfunctions that could result in financial losses or other damages.
Zurich Insurance, who have created a whitepaper called ‘Artificial Intelligence And Algorithmic Liability’, noted that the “AI risk” is defined as arising from the use of data analytics and cognitive technology-based software algorithms in automated and semi-automated decision-making. Where the algorithms fail to perform as expected, the result can be losses from property damage, business interruption, personal injury, professional liability and cyber exposures.
To give a simple yet reasonable example, if an AI system generates a design that is not suitable or infringes on someone else’s intellectual property, the designer could face legal action from their claims in addition to potentially hefty fines.
There are two issues in this scenario:
The first is that there is no certainty about where the error or omission exists. Is it the creators of the AI and the algorithms that are at fault or is it the designer for not having adequate mechanisms for checking the work and rectifying any errors?
From a claims perspective, it is difficult to determine how this issue would be approached by both Insurers and the Courts. As outlined by Zurich in their whitepaper, the insurance implications will take some time to determine “… because of a lack of loss experience data and models that can estimate the frequency and severity of potential losses …”.
The first issue leads into the second, which relates to the cover position.
It is likely that insurance coverage held by the designer relates to the provision of their professional services. Whilst there are unlikely to be any terms within those policies that would prohibit the use of an AI system, has the designer provided a professional service if they have used AI to complete the project?
As it is likely that the designer will ultimately retain a duty to inspect and approve any design work before it is released, if the designs are in error there would be an expectation that a further related claim could be brought against the providers of AI. But, if the terms of any contract with the AI provider does not provide any route for a claim (because they accept no liability as mentioned in their agreement), the position of Insurers has arguably been prejudiced.
Whilst it is difficult to predict how Insurers would react in these circumstances, it is possible that they would look to restrict or remove any cover offered in respect of any designs produced by AI Systems. As such, there is a great deal of uncertainty regarding the future landscape of the insurance sector as these new technologies continue to emerge.
Intellectual Property Issues
Another important consideration when using AI in design is intellectual property (IP) rights. AI-generated designs may incorporate elements or features that are protected by copyright, trademark, or patent laws. As such, designers and companies need to understand the potential IP issues associated with the use of AI in design.
Again, for simplicity… an AI system may generate a design that incorporates a copyrighted image or trademarked logo without permission. This could result in legal action against the designer and their client for copyright or trademark infringement, which could incur defence costs.
Whilst the AI creator is likely at fault, this is not certain. Therefore, it would be prudent to avoid these risks, but if you are using AI, we suggest being vigilant in checking the legality of any elements used in AI-generated designs.
In addition, designers and companies may also need to consider issues related to the ownership and licensing of AI-generated designs. In some cases, the copyright and other IP rights in an AI-generated design may be owned by the developer of the AI system, rather than the designer or their client.
As such, designers and companies need to clarify ownership and licensing rights upfront to avoid any disputes or legal action down the line.
Whilst AI has the potential to revolutionise the world of design, it also presents a range of unanswered questions about insurance, intellectual property, and risk management, which all need to be carefully considered.
The good thing is it is generating a lot of debate amongst regulators and technology companies. But until these questions are answered, it is important that you develop clear policies and procedures for using AI. These policies should include guidelines for testing and validating AI systems, as well as protocols for reviewing designs and responding to any errors or malfunctions that may occur.