Ireland – The AI Island and the need for legislation

The Artificial Intelligence industry is growing exponentially across the world each year. It is weaving its way into almost every aspect of our lives, from travel to online connectivity, to music and video streaming and to customer service in online retailers such as Amazon and eBay.[1] Many nations around the world are vying to be the country to win the “AI race”, the United States, China, Germany and the United Kingdom, to name but a few. However, Ireland, despite its smaller size than the aforementioned, has become home to a vibrant “AI ecosystem”[2] thanks to the highly educated workforce, low corporate tax rate, Research and Development tax credits of 25% and a 6.25% preferential tax rate on income derived from qualifying intellectual property.[3] The work of agencies such as the IDA, Enterprise Ireland and ICTSkillsnet, among many others, has resulted in the creation of this flourishing AI ecosystem and Ireland has now garnered the reputation of an AI Island.[4] Some of the leaders in Artificial Intelligence in Ireland include: IBM, Nuritas, Movidius, Accenture, Veritas and Xilinx. [5]However, despite Ireland being an AI hub in Europe, there is no legislation, which covers the use and creation of Artificial Intelligence specifically.[6]

Artificial Intelligence is currently governed by a combination of existing laws and legislation in Ireland, for example Copyright Law, Tort Law, Data Protection Laws and Contract Law. A key example of this is in terms of Copyright issues. Under the Copyright and Related Rights Act 2000, a legal person – an individual, a company or a partnership, can only own copyright. Therefore, in order for work to attract copyright protection, a legal person must create it and so the issue arises in the situation where a machine, through artificial intelligence, creates work. In this instance, can the machine be the owner of the work? The answer here is no, as a machine is not a legal person under Irish law. As a result, the ownership of the work created by Artificial Intelligence must be decided through contractual agreements in order to attract Copyright protection. [7] Therefore, we see how this is not a straightforward application of the law but rather a mélange of various legislation. 

Another problem to be considered whether a duty of care arises in relation to the use of AI. The Supreme Court judgement in Glencar Exploration plc v Mayo County Council [8] embraced the broader view of a duty of care, which followed the approach of the English courts. It was established that for a claim of negligence to be successful, three requirements must be shown; the existence of a duty of care; the breach of that duty; and proof that the breach was the cause of the injury which is the subject of the claim.[9] Precedence shows that a manufacturer can be liable in negligence for a defective product, which caused injury if these requirements are established.[10] However, the issue becomes far more complex when Artificial Intelligence is involved. If a product of Artificial Intelligence gives options to a user of the product and the option chosen results in damage to the user, it may be difficult to definitively link negligence and liability back to the manufacturer. A similar question can be raised in relation to the tort of ‘escape and strict liability in respect of Artificial Intelligence. An interesting analogy can be drawn with the Control of Dogs Act 1986[11], under which  strict liability is imposed on owners for injury or damage caused by their dogs. Parallels have been drawn between vicious dogs and products and services of artificial intelligence. As AI continues to develop, will common law aspects of strict liability be able to apply to artificial intelligence or will it be necessary to develop legislation to deal with these issues specifically? [12]

Another consideration to be taken into account is that of Data Protection and Cyber security and the implications these have for Artificial Intelligence. Data protection in Ireland is governed by the General Data Protection Regulation[13] and is transposed into Irish law by the Data Protection Act 2018[14]. Article 35 of the GDPR states that where processing which uses new technology, will likely result in jeopardising the rights and freedom of people, the controller shall, prior to the processing, carry out an assessment of the impact of the processing operations on the protection of personal data.[15]Furthermore, article 25 concerns privacy by design and default and obliges controllers to ensure that applications, specifically those involving AI, are designed to incorporate data protection principles from inception in order to meet the data protection requirements.[16] However, once again, this legislation is not specific to artificial intelligence and so as AI continues to develop and advance, the application of existing legislation and laws will become more complex and more problematic. As a result, this may lead to loopholes being manipulated in areas of legislation, which do not specifically relate to artificial intelligence. 

Other issues arise in relation to the use of Artificial Intelligence in Healthcare due to the fact that in order to establish which liability framework can be applied; it is necessary to determine when  the AI used is considered to be a product or a service.[17] Similarly, it is predicted that AI will revolutionise the financial services industry and while there is currently no prohibition on the use of AI by Financial Service Providers, it will be crucial that the necessary controls are in place to provide and ensure sufficient and appropriate oversight.[18]

The examples listed above are just a few of the sectors affected by Artificial Intelligence, but the reality is that AI is seeping into every aspect of our lives. AI is developing and improving at an unprecedented speed and while this is a positive thing, it is imperative that it is accurately and sufficiently regulated. The law tends to lag behind developments in technology and seems to prefer to play a game of catch-up rather than get ahead of curve. In relation to regulations and safeguards for the use of artificial intelligence, there is no time for a learning curve. In a field as new and revolutionary as AI, mistakes are inevitable and so rather than learning from these inescapable mistakes, it would be expedient for the law in Ireland to pre-empt them through specific legislation. Without specific legislation, systems, products and services, such as algorithms which use AI, are more likely to result in inaccuracies as companies have less incentives to ensure safety measures are put in place and to assure quality. [19]

An important advancement in legislation for Artificial Intelligence came about in April 2021 when the European Commission published its proposal for a Regulation on Artificial Intelligence.[20] The regulations introduce a comprehensive regulatory framework for Artificial Intelligence in the EU and aim to establish a legal framework that provides the necessary certainty to enable and facilitate innovation in AI.[21] The regulations will also establish the required safeguards for fundamental rights and ensure the AI is used and implemented safely and accurately.[22] The proposal is a combination of a legal framework for Artificial Intelligence and a Coordinated Plan with Member States. As Margrethe Vestager, the Executive Vice-President for a Europe fit for the Digital Age, commented: “On Artificial Intelligence, trust is a must, not a nice to have.” [23] This is the attitude, which must constantly be borne in mind when thinking of artificial intelligence. Artificial Intelligence is undoubtedly a hugely advantageous development in technology in society, however, in order for it to be used most effectively, we must be able to trust it and its implementation in our lives. Vestager continued to note that through the Commissions regulations, the EU is “spearheading the development”[24] of Artificial Intelligence so that it can be certain that AI can be trusted and so that the EU can “pave the way to ethical technology”[25]. 

In light of the new proposal for regulations by the Commission, it would be an opportune time for Ireland to follow in their footsteps, as the “AI Island”, by developing our own legislation to specifically regulate Artificial Intelligence. This will not only help to ensure the necessary safeguards are in place, but will also aid in the promotion of Ireland as a hub for artificial intelligence. 

[1] Mason Hayes and Curran, ‘Artificial Intelligence – Overview’ (June 2019) <https://www.mhc.ie/uploads/MHC1324_Artificial_Intelligence_Booklet_%28Ezines_A5_148x210%29_WEB.pdf>

[2] Kevin Harnett and Claire Morrissey, ‘AI, Machine Learning, Big Data Laws and Regulations 2021- Ireland’, (Global Legal Insights, 2021) https://www.globallegalinsights.com/practice-areas/ai-machine-learning-and-big-data-laws-and-regulations/ireland

[3] Ibid 2. 

[4] Ibid 1. 

[5] IDA Ireland, ‘Artificial Intelligence – Ireland AI Island’ https://www.idaireland.com/newsroom/publications/artificial-intelligence

[6] Victor Timon, ‘Ireland: Artificial Intelligence Comparative Guide’, (Mondaq, 6 January 2022) https://www.mondaq.com/ireland/technology/1059764/artificial-intelligence-comparative-guide

[7] ibid 5. 

[8] Glencar Exploration plc v Mayo County Council [2002] IESC 64.

[9] ibid 7. 

[10] Ibid 5.

[11] Control of Dogs Act 1986

[12] Ibid 5. 

[13] EU General Data Protection Regulation (GDPR): Regulation (EU) 2016/679

[14] Data Protection Act 2018

[15] ibid 12. 

[16] Ibid 12. 

[17] Ibid 1. 

[18] Ibid 1. 

[19] Jascha Galaski, ‘AI Regulation: Present Situation and Future Possibilities’, (Liberties, 8 September 2021) < https://www.liberties.eu/en/stories/ai-regulation/43740. 

[20] Press Release, ‘Europe fit for the Digital Age: Commission proposes new rules and actions for excellence and trust in Artificial Intelligence’ (European Commission, 21 April 2021) https://ec.europa.eu/commission/presscorner/detail/en/ip_21_1682

[21] Arthur Cox, ‘The EU’s new Regulation on Artificial Intelligence’ (Briefing, 6 May 2021) <https://www.arthurcox.com/knowledge/the-eus-new-regulation-on-artificial-intelligence/> 

[22] Ibid 20.

[23] Ibid 19.

[24] Ibid 19. 

[25] Ibid 19.