By Rakhi Rashmi, Barrister at 10 Kings Bench Walk Chambers

 Introduction

There is a widespread recognition that Artificial intelligence [AI] can swiftly and dramatically transform human reality in all sectors. Along with having the power to transform most sectors, the potential risk associated with AI systems is also of concern. The benefit of AI in the legal industry could be enormous as using AI technologies could help in all stages, from negotiations to litigation, resolving conflicts to court judgments, and conducting preliminary legal research. Notwithstanding the enormous benefit of its application in the legal sector, the potential risk is also of concern. The risk associated with the usage of AI in the legal industry can be understood by the example of a recent case of Mata v. Avianca in 2023 in which the usage of ChatGPT had created fabricated citations. It is of concern that the trend of similar cases is on the rise [Parsons 2024]. Similarly, all other sectors are equally skeptical about increasing their dependability on AI without robust regulations that can mitigate the associated risk and guarantee safety and privacy. Due to the increased risk concerns, global governments have initiated regulating AI to have the framework of laws and regulations to ward off potential associated risks by appropriately balancing growth and safeguarding fundamental rights.

The recent regulatory developments by the European Union (EU) and the United States, with a white paper along with a private member AI Bill in the United Kingdom, show the desire of global governments to react to the dangers of AI and its usage [Haie et al. 2024]. More recently, the G7 leaders pledged to remain dedicated to setting AI standards that balance their shared democratic values.

Considering the growing focus in recent years on the need to regulate AI technology, it is interesting to analyse the trend of divergent global AI regulatory initiatives and their approaches, which may shape the future AI regulatory landscape.

AI regulations and EU

The UK government’s efforts in regulating AI can be traced back to 2018, when the government published an AI Regulation Policy paper setting out the UK’s vision for the future ‘pro-innovation’ and context-specific AI regulatory regime. In 2022, a UK. government approach introduced “A pro-innovation approach for Regulating AI” [White Paper, 2023], which states that the cross-sectorial principles for various sectors ensure AI usage for safety and AI governance. Safety, security, transparency, fairness, accountability, and management are mainly used in this framework.

There has been an effort to bring top global scientific minds together by way of organizing the AI Safety Summit held at Bletchley Park in November 2023 on the Safety of Advanced AI which was an effort to globally develop an interchange of information of the opportunities and risks posed by rapid advancements in AI and tackle one of humanity’s greatest challenges [Szczepanski and Killmayer 2024].

The UK White Paper has identified the potential risks associated with the increasing application of AI, which are Risks to human rights, safety, fairness, privacy, societal well-being, and security, but has proposed that sector-specific regulators Certainly, here is the revised text: “We expect” to interpret and apply the AI framework within their respective domains[White Paper, 2023]. The sector-specific regulators will have the responsibility of having regard to its principles and governance of actors across the life cycle of AI systems.

UK and EU AI regulatory approach

The EU AI Act which is a horizontal regulation that has harmonized legislation imposes legislative obligations at all sectors and all stages of the AI life cycle right from training to testing to validation; to assessments; risk management systems; and post-market monitoring [EU AI Act, 2024].

Now considering the EU AI regulatory effort the EU has already enacted its EU AI Act [which was passed in the EU parliament on March 2024 and will be fully applicable around June 2026] which encapsulates a prescriptive legislative framework by adopting a risk-based approach to regulate the development and usage of AI across all member states. Compared to the UK’s approach, the EU has a divergent approach. The European Union AI Act provides AI developers and deployers with precise requirements and obligations regarding specific uses of AI by categorizing four levels of risk—unacceptable, high, limited, and minimal [EU AI Act, 2024].

Global AI regulatory landscape trend

In the U.S., numerous efforts have been made to develop guidelines and frameworks to regulate AI Mostly, regulations are currently being examined at both the federal and state levels [central and regional]. Although there is no legislation in place on the Federal level yet, few states have already introduced AI legislation aimed at regulating AI, particularly around privacy and accountability, and has a more case-to-case basis approach.

Now, about the other global efforts in countries such as Brazil, China, Japan, India, Canada, Switzerland, and Australia as all are in action having AI governance or the Act for AI Regulations via sector and horizontal approach parallel to EU AI Act [Kohn and Pieper 2023]. In China, there have been many relevant regulations in the last few years, including data protection laws, ethical codes for AI, etc. Still, the major highlight of the Chinese AI laws is that they tend to regulate private companies and not government-owned ones [Kohn & Pieper, 2023]. Switzerland’s AI approach parallels the UK through sectoral-based AI regulation and not having standalone legislation[Kohn & Pieper, 2023]. Such AI regulations can be seen as a more business-friendly approach to AI regulation not having excessive restrictions in Switzerland.

Conclusion

Although there is an increasing trend towards regulating AI it is challenging to predict a global AI regulatory landscape at this time. The UK has taken a non-statutory and sector-based AI regulation approach. In the UK, there still needs to be more sector-specific guidance which are published to date. Relevant guidance may soon be in the pipeline for the relevant sectoral regulators to regulate the sector-specific AI issues. The UK’s non-statutory principle-based framework makes it challenging to ensure that AI systems comply with the AI Act [Pearce 2024]. However, we need to wait to say anything for sure.

At the same time, it can be said for sure that it remains an uphill task for the UK government to have a balanced framework on a sectoral basis that can promote innovation from design to research to training and development to sales while fostering an ethical and responsible atmosphere.

The closely connected regions, such as the EU and the U.S., work towards harmonizing their legislation with a network of new and established regulators, including a central European AI Board, etc, for AI in each Member State. In contrast, the UK relies upon existing sectoral regulators, which include the Medicines and Healthcare Products Regulatory Agency, the Information Commissioner’s Office, the Equality and Human Rights Commission, etc. The UK approach adapts the fundamentals to the specific context in which AI is deployed by using the domain-specific expertise of regulators[White Paper, 2023]. Therefore, the UK must work on potential regulatory gaps while adopting a sectoral regulatory system to facilitate easier global collaboration and trade for innovators. The UK Government knows that regulatory coordination is still essential given its non-statutory context-specific approach. The risk is that in the absence of a centralized, single regulator, UK regulators will diverge in approach. Although the white paper contains guidance for regulators to mitigate risks while implementing the principles with recommendations for central monitoring and evaluation, its success on the implementation level still seems elusive and unpredictable.

Considering the global trend and efforts to regulate AI, there is no clarity on the Global AI regulatory landscape at this time. However, there has been a growing trend to lead in the advancement of AI while minimizing risks. The UK and other nations are on a mission to balance innovation, safety, and ethical application of AI.

Global regulators must critically examine the fundamental aim of AI regulation, which is to uphold ethics and responsibility while fostering equality, privacy, and AI safety.  A reasonable regulatory framework that preserves human rights and encourages innovation will be crucial in ensuring that artificial intelligence (AI) advances to the advantage of all while preserving trust and safety.

Rakhi Rashmi, Barrister at 10 Kings Bench Walk Chambers

References:

1. Black, J., & Murray, A. D. (2019) Regulating AI and machine learning: Setting the regulatory agenda. https://core.ac.uk/download/275552623.pdf
2. https://web.stanford.edu/class/sts175/NewFiles/Negroponte.%20Being%20Digital.pdf [Accessed 19 August 2024].
3. https://books.google.co.uk/books/about/Republic_com.html?id=O7AG9TxDJdgC [Accessed 19 August 2024].
4. https://ir.lawnet.fordham.edu/faculty_scholarship/42 [Accessed 19 August 2024].
5. https://chicagounbound.uchicago.edu/uclf/vol1996/iss1/7/ [Accessed 19 August 2024].
6. https://cyber.harvard.edu/works/lessig/LNC_Q_D2.PDF [Accessed 19 August 2024].
7. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=535 [Accessed 19 August 2024].
8. The King’s Speech sets out the new Labor government’s proposed laws and its plans for the upcoming parliamentary term. Prime Minister’s Office and His Majesty King Charles III (2024); https://www.gov.uk/government/speeches/the-kings-speech-2024 [Accessed 19 August 2024].
9. A. Prinsley, O. Yaros, R. Randall, O. Hajda, E. Hepworth ‘ UK’s Approach To Regulating The Use Of Artificial Intelligence’ Mayer Brown (2023); https://www.mayerbrown.com/fr/insights/publications/2023/07/uks-approach-to-regulating-the-use-of-artificial-intelligence [Accessed 19 August 2024].
11. G. Haie, T. Cohen, A. Golodny, M. Shenk, M. Avramidou, E. Goodwin & V. Arethusa. “A Comparative Analysis of the EU, U.S. and UK Approaches to AI Regulation.” Steptoe, 2024; https://www.steptoe.com/en/news-publications/steptechtoe-blog/a-comparative-analysis-of-the-eu-us-and-uk-approaches-to-ai-regulation.html [Accessed 19 August 2024].
12. Kohn & M. F.U. Pieper. “AI Regulations around the World.” Taylor Wessing, 2023; https://webadmin.taylorwessing.com/en/interface/2023/ai—are-we-getting-the-balance-between-regulation-and-innovation-right/ai-regulation-around-the-world [Accessed 19 August 2024].
13. Parsons “AI in 2024: A Year of Crossroads and Decisions.” Georgia State University College of Law, 2024; https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4821719 [Accessed 19 August 2024].
14. Pearce “Regulating AI in the EU and the UK – A Legal View.” Foundation Blog, 2024; https://www.foundation.org.uk/Blog/2024/Regulating-AI-in-the-EU-and-the-UK-a-legal-view [Accessed 19 August 2024].
15. Ponomarov “Global AI Regulations Tracker: Europe, Americas & Asia-Pacific Overview.” Legal Nodes, 2024; https://legalnodes.com/article/global-ai-regulations-tracker [Accessed 19 August 2024].
16. “Regulation of Artificial Intelligence.” Wikipedia, 2020; https://en.wikipedia.org/wiki/Regulation_of_artificial_intelligence [Accessed 19 August 2024].
17. Roberts, A. Babuta, J. Morley, C. Thomas, M. Taddeo, and L. Floridi. Artificial Intelligence Regulation in the United Kingdom: A Path to Good Governance and Global Leadership? by Huw Roberts, Alexander Babuta, Jessica Morley, Christopher Thomas, Mariarosaria Taddeo, Luciano Floridi :: SSRN. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4209504[Accessed 19 August 2024].
18. Scali, P.E. Froge, and J. Dunn. “Europe: UK vs EU Approach to Regulating AI” DataGuidance, 2022; https://www.dataguidance.com/opinion/europe-uk-vs-eu-approach-regulating-ai [Accessed 19 August 2024].
19. Szczepanski & G.L. Killmayer. “The United Kingdom and Artificial Intelligence.” European Parliamentary Research Service (April 2024); https://www.europarl.europa.eu/thinktank/en/document/EPRS_ATA(2024)762285 [Accessed 19 August 2024].
20. Tongia “Why Regulating AI Can Be Surprisingly Straightforward When Teamed with Eternal Vigilance.” European Sting, 2024; https://europeansting.com/2024/05/29/why-regulating-ai-can-be-surprisingly-straightforward-when-teamed-with-eternal-vigilance/ [Accessed 19 August 2024].
21. Hickman, J. Rennie, and A. Jha. “AI Watch: Global Regulatory Tracker – United Kingdom.” White & Case; https://www.whitecase.com/insight-our-thinking/ai-watch-global-regulatory-tracker-united-kingdom [Accessed 19 August 2024].
22. Tobin ‘Artificial Intelligence (Regulation) Bill [H.L.]’ HL Bill 11 of 2023–24; https://researchbriefings.files.parliament.uk/documents/LLN-2024-0016/LLN-2024-0016.pdf [Accessed 19 August 2024].
23. Gallo & S. Nair. “The UK’s Framework for AI Regulation.” Deloitte, 2024; https://www.deloitte.com/uk/en/Industries/financial-services/blogs/the-uks-framework-for-ai-regulation.html [Accessed 19 August 2024].
24. Benizri, A. Evers, S.T. Mercer, and A.A. Jessani. “A Comparative Perspective on AI Regulation.” Lawfaremedia, 2023; https://www.lawfaremedia.org/article/a-comparative-perspective-on-ai-regulation [Accessed 19 August 2024].
25. Hickman, Dr. S. Lorenz, Dr. C. Teetzmann, A. Jha. “Long Awaited EU AI Act Becomes Law After Publication in the EU’s Official Journal.” White & Case; https://www.whitecase.com/insight-alert/long-awaited-eu-ai-act-becomes-law-after-publication-eus-official-journal [Accessed 19 August 2024].
26. United Kingdom : Countries agree to safe and responsible development of frontier AI in landmark Bletchley Declaration. MENA Report, (2023).
27. UK vs EU Approach to Regulating AI: From One Extreme to Another? – Connect On Tech. https://www.connectontech.com/uk-vs-eu-approach-to-regulating-ai-from-one-extreme-to-another/
28. Swiss FADP Compliance for Life Science Companies. https://www.mydata-trust.com/2023/02/13/swiss-fadp-life-science-companies

Legislations/ Regulations

26. White Paper Policy paper: ‘A pro-innovation approach to AI regulation’ GOV. UK Updated 3 August 2023, p. 45, 46,11, 15, 25, 78; https://www.gov.uk/government/consultations/ai-regulation-a-pro-innovation-approach-policy-proposals/outcome/a-pro-innovation-approach-to-ai-regulation-government-response [Accessed 19 August 2024].
27. European Commission, ‘AI Act’, 6 March 2024; https://www.europarl.europa.eu/doceo/document/TA-9-2024-0138_EN.html [Accessed 19 August 2024].
28. Mata v. Avianca, Inc., 678 F. Supp. 3d 443 – Dist. Court, SD New York 2023