Artificial Intelligence Update: AI in Commercial Contracts - What You Need to Know
21 May 2019
Significant attention has been drawn to AI in the context of headline areas such as data protection and cyber security. However, one of its most practical applications to date in the legal industry has been for commercial contracting. In particular, the drafting and negotiating of commercial contracts for AI-based products and/or services has given rise to specific consideration given the nature of the technology and its potential impact within the business. We focus on the areas of disclaimers/insurance, confidentiality and audit rights and AI and the legal liability problem.
Disclaimers and insurance policies
This issue should be considered carefully in any contract for an AI service. An AI supplier will generally seek to provide the AI service with limited guarantees related to the standard of service and care should be taken to ensure that the supplier does not exclude any warranties relating to the quality of the AI or its service availability. As AI evolves, there is a greater potential for service outages; and the general impact this might have on the wider business should be taken into account, particularly around support services, etc. Customers receiving AI services should also review and negotiate liability clauses with caution, as it is likely that a supplier will have capped or excluded liability for the unavailability of the AI service.
It is also prudent for customers to explore both their own existing insurance policies and also the insurance coverage which their supplier has in place. We have seen a number of instances over recent years where new technologies or risks around the use of AI technology are not automatically captured within existing policies. This is an important consideration and should be explored with the company’s insurance brokers.
Confidentiality or auditing rights
The standard auditing clause in commercial contracts will take on a new dimension in light of AI. This is because the work typically carried out by personnel will be carried out through AI systems instead. Many suppliers will be hesitant of permitting wide audit rights around their AI technology and will usually resist anything that might impact upon their proprietary technology. As a result, it will be important to ensure that existing audit clauses and mechanics used in standard supplier agreements are examined to ensure they are flexible enough to cover verification of AI-produced results where this is required for internal audit purposes.
Confidentiality issues are another area worth exploring, particularly during the input of business sensitive information within the system. Difficulties can arise for customers to ensure that once an AI agreement has been terminated the information will be deleted from AI systems which are generally, by their nature, built to use and learn from the information which it is processing. In addition, the issue of migration to new service providers can also be potentially problematic. This is because of the proprietary nature of many AI systems and the reluctance of many suppliers to provide any information which might compromise their ‘black-box’ technology.
AI and the legal liability problem
AI is not responsible for its own actions because it is not a legal person. As a result, the question then arises as to who bears legal responsibility? Similarly, under the laws of negligence, questions arise around duty of care and other related concepts. This is still somewhat unknown territory and is one of the most debated legal aspects of AI as a whole. Accordingly, it is likely that we will begin to see a new approach to liability clauses and how they are constructed over the coming years in order to address these difficult issues.
AI is still evolving and we are likely to see a continued diversity in terms of how businesses use and implement the technology into their existing business models. While AI specific commercial contracts are still very much a work in progress, it is clear that contracts being entered into for AI-based products and/or services will require additional consideration prior to drafting. Otherwise there is the potential for contracts to be impractical, not fit for purpose and overall being disruptive to the operations of the business. It is, therefore, possible that these negative factors could negate much of the benefit of moving to AI.
The content of this article is provided for information purposes only and does not constitute legal or other advice.