The liability of robotic and autonomous systems
Recommendations for law, policy and ethics of AI for the United Arab Emirates (UAE)
This is my final article in this series; a great way to start the New Year! After discussing the liability of autonomous systems under the UAE law, identifying gaps and comparing to other regimes, this article, in a way, is the practical conclusion of my research and so I will discuss my recommendations for law, policy and ethics for the United Arab Emirates (UAE).
1 — Legislative changes recommendations
We have a tendency to regulate any new matter in a civil law regime. Whether a decree or a law, we issue regulations with the purpose of covering any potential gaps or uncertainty. With technology, it may not be practical to follow this approach as it continues to evolve, with new developments introduced on a regular basis.
This is why the concept of tech neutrality was introduced some time ago as a generally acceptable regulatory approach. The concept means that “legislation is drafted in a manner that is not bound to any specific technological form or method. The objective is two-fold. First, to enable legislation that lasts the test of time and does not become outdated when technology develops. Second, to treat different technological solutions equally and without inadvertently granting unfair advantage to certain solutions while discriminating others.”
The technology adopted in the UAE (as is the case in most of the world) have not yet reached a fully autonomous state. Consequently, in its current form, the UAE legislative regime is broad enough to cover the liability of these autonomous systems. The civil and criminal laws, in addition to the consumer protection law as discussed in previous posts, offer enough protection to injured persons. However, in preparation for full autonomy, I recommend minor changes to the current UAE legislative regime.
Insurance
Insurance is one of the solutions contemplated in western jurisdictions to cover any legislative gap. As a quick reminder, the gap identified in fully autonomous systems stems from their continual learning from their environment and that “defects” may no longer be attributed to the manufacturer of the system.
Therefore, one of the first adjustments that can be made is to make sure an injured party receives compensation for the damages he or she suffered, through insurance. This avoids the need — at least for now — to go into detailed debates about which party is liable in case of injury and wait for the courts to decide on matters.
For example, a practical, direct and quick approach would be to use the existing regulations regarding the Unifying Motor Vehicle Insurance Policies issued by the Federal Insurance Authority. The current regulations make it mandatory for each vehicle operating in the UAE to have insurance. An autonomous vehicle should be no different. This would be a catch-all solution and only minor tweaks may be needed for the existing regulations.
What about other autonomous systems outside the automotive industry? Insurance companies are going to have to shift from a ““detect and repair” [approach] to “predict and prepare”.” Cybersecurity insurance is already one of the fastest growing insurance markets for small and medium sized businesses.
Most technology multinationals have third party liability insurance as part of their worldwide policies. The Federal Insurance Authority could mandate specific software or machine learning local insurance to cover for any gaps, especially in the medical field.
Alternatively, soft law can be adopted as an interim measure before actual legislation is implemented. I will discuss this more below while focusing on ethics as well.
2 — Soft law recommendation and ethics
What does soft law mean? In the AI context, it is simply issuing a set of principles that will govern the use and deployment of AI.
In the UAE, there are the principles issued by Smart Dubai, the “Dubai AI principles.” They can be considered as a basis for any countrywide framework.
These principles may not have the same force of law but they will serve as guidance when later implementing specific regulations and they may become binding if widely adopted.
In my opinion, a good example would be to issue a set of principles similar to Singapore’s Model AI Governance Framework. The principles do not have to be as extensive as those adopted in Singapore but a good start will be to cover:
- liability
- intellectual property
- safety
- ethics of AI
Speaking of ethics, there are many papers and articles discussing the ethics of AI. AI systems are basically the products of their designers and so many ethical issues that are found in our society when it comes to human interaction can be replicated in AI. Issues include AI bias, privacy, transparency and accountability. Many of the big players have developed principles to guide their own work in AI, including Microsoft’s “Responsible AI” and others.
At HPE, we are also working on our principles, and have tackled consciousness of AI and the questions around it.
The Berkman Klein Center for Internet and Society at Harvard University recently issued “Principled Artificial Intelligence: Mapping Consensus in Ethical and Rights-based Approaches to Principles for AI.” I highly recommend reading the paper. They identified eight mains themes among AI principles, distributing them as follows:
- Privacy (8 principles)
- Accountability (10 principles)
- Safety and security (4 principles)
- Transparency and explainability (8 principles)
- Fairness and non-discrimination (6 principles)
- Human control of technology (3 principles)
- Professional responsibility (5 principles)
- Promotion of human values (3 principles)
As part of the soft law recommendation for the UAE AI Principles, ethics must play a big role. The principles that should be adopted should focus mainly on the design and the implementation phases, making sure autonomous systems are safe and secure and respect the privacy of individuals.
In a society like the UAE, where this subject has been lately emphasised, it becomes a must. For example, we have seen enforcement of the Cyber Crime Law in social media and the latest amended consumer protection law issued on 10 November 2020 discussing privacy of consumer data. Any AI system must respect the culture and norms of the UAE and so the subject of ethics and privacy in AI is extremely relevant and something which I may discuss in future articles.
In conclusion, through insurance regulation and soft law, we can make immediate changes to cover any potential gap in the legislation and set a framework for AI deployment in the UAE. In order to do that, collaboration between the various government departments and the private sector is essential and through small initiatives, lobbying and discussions about the topic in conferences and articles such as this one, we can raise more awareness and start the process of change.
Tarek Nakkach
Regional Legal Counsel — UKIMESA
Hewlett Packard Enterprise
and AAIP Programme Fellow
Contribute to the strategic development of the Programme as a Fellow.
assuring-autonomy@york.ac.uk
www.york.ac.uk/assuring-autonomy
This post is intended to be used for information purposes only and not as legal advice.
For the purpose of this post, the term autonomous systems covers robotics and Artificial Intelligence (AI) software or algorithms; any system that could function without human intervention or control. The focus from a legal point of view is on the term “autonomous”.
