Artificial intelligence, or ‘AI’, has become a rather topical issue, envisaged to bring about wide-ranging economic and societal effects across the board, impacting various sectors such as healthcare, finance, mobility, agriculture and climate change mitigation, among others.
The recent advancement of digital technologies has brought AI closer to us, allowing AI to go beyond the limits set by humans – which is essentially what makes the application of AI to our lives so exciting. While this fast-evolving family of technologies is expected to bring about numerous benefits, certain risks and negative impacts may also be expected, thus making proper regulation and education on AI ever so important.
The EU’s proposal for an ‘Artificial Intelligence Act’
On April 24, 2021, the European Commission announced its proposal for a regulation on the laying down of harmonised rules on AI. This proposal constitutes the first-ever legal framework on AI, addressing certain risks, and putting the EU at the forefront of the global scene. Fostering excellence in AI will definitely boost Europe’s, and Malta’s, potential to compete globally, and the setting up of a sound legal framework is definitely the place to start.
In Article 3 of the proposed ‘AI Act’, a definition of an “artificial intelligence system” has been set out, which inter alia establishes that an AI system refers to software which can generate outputs such as content, predictions, recommendations, or decisions influencing the environments they interact in, in line with a set of human-defined objectives. It is being submitted that this definition is a somewhat restrictive one, limiting AI systems solely to software forms, and excluding other possible forms, such as biological forms for example.
The term ‘artificial intelligence’ is also misleading in itself since in reality, the technology has little to do with our human intelligence. Additionally, since there is currently no uniform definition of the term ‘intelligence’, how can legislators possibly define ‘artificial intelligence’, and subsequently structure a comprehensive and binding legal framework in this regard?
The proposed Act also classifies AI systems into three different categories, depending on the level of risk involved in that system. A sort of ‘traffic light’ system is created, whereby those systems carrying an ‘unacceptable risk’ will be prohibited, those carrying a ‘high risk’ will be subject to harsher regulations (particularly when it comes to security and accountability matters), while those carrying a ‘low or minimal risk’ will only be subject to transparency obligations.
Malta’s AI legislative framework
Currently, there is no legislative framework regulating AI in Malta. Having said this, several measures and initiatives have been taken in this regard, primarily by the national ‘Malta.AI Taskforce’, which published A Strategy and Vision for Artificial Intelligence in Malta 2030, and Malta’s Ethical AI Framework.
The term ‘artificial intelligence’ is also misleading in itself since in reality, the technology has little to do with our human intelligence- Nina Fauser
These two key documents, which were published in April and October 2019 respectively, despite not being legally binding, nonetheless contribute to Malta’s regulatory framework on AI, in line with the AI Ethics Guidelines issued by the EU Commission’s High-Level Expert Group on AI.
It will be interesting to see how potential future local legislation on AI will be impacted by the introduction and implementation of the EU AI Act, which could potentially be used as a basis for drafting Malta’s legislation. Considerations relating to risk, as well as ethical considerations, will also surely be included.
Data protection and privacy concerns in AI
With the rapid development of AI technologies, there is an inevitable increase in the level of risk of personal information being used in ways which could potentially infringe an individual’s data protection and privacy rights.
Interestingly, in terms of the General Data Protection Regulation (GDPR), when a decision is taken solely in an automated manner, such decision must allow for human review if it significantly affects the data subject. Moreover, the data subject has a right to be informed regarding the manner in which such decision was taken. This evidently imposes quite a significant limitation on the automation of AI systems and on companies developing such systems.
In order to structure a comprehensive and sound legislative framework on AI, a balance ought to be achieved between safeguarding the use of one’s personal data on the one hand, and establishing the right set of rules which do not impose too many restrictions which hinder the development of automation and AI technologies on the other hand.
In conclusion, it is recognised that AI technologies constitute an extremely powerful tool, which have the potential of revolutionising the world and the manner in which things are currently being done. Having said this, due to its immense potential, it is imperative that this tool is used cautiously, and that a proper legislative framework is put in place without delay.
AI is there to enrich our basic human rights and, therefore, we must agree on a governance framework to ensure the effective protection of these rights.
Notwithstanding the extensive research being undertaken in this field, further detailed research is yet to be explored, which must work hand in hand with the setting up of a comprehensive legislative framework in this regard.
The ‘Prof. Joseph A Micallef Essay Prize in Law’ competition was held for the first time under the auspices of the Law Faculty of the University of Malta. The competition provides the faculty’s undergraduate students the opportunity to share their views on a topic of their choice, with the aim of promoting advocacy and research, and fostering the articulation of legal thought.
The competition was organised thanks to the generosity of the family of the late Prof. Joseph A Micallef. On the recommendation of foreign leading academics, the submission by Nina Fauser on the future of Artificial Intelligence was selected as the winning submission of the first edition of the Essay Prize. The above is a summarised version of the submission.