Emerging AI technologies and the potentially existential threats they pose to human society are a hot topic at the moment. These technologies have the potential to improve many aspects of human society, but the existential risks they pose have been well-publicized.
Image Credit: Ivan Marc/Shutterstock.com
Seeking to get ahead of the curve, so to speak, on AI, the European Commission has proposed the EU AI Act, the first comprehensive law in the world related to AI and how to regulate it. This article will explore what the EU AI Act is, its provisions and objectives, and the impact it will have on both the nascent AI industry and wider society.
Introduction to the EU AI Act
AI is a tool that has undoubtedly and profoundly revolutionized several industries. It can provide benefits such as reducing operational costs, improving decision-making, quickly analyzing big data sets, automating repetitive tasks, and technologies such as machine learning can produce more efficient, almost human-like tools.
However, AI is not without its potential drawbacks. Generative AI has the potential to be highly disruptive to the creative industries; deepfakes of political figures and celebrities could spread misinformation on a grand scale, and there is even a well-founded fear that sophisticated AI tools could be used in cybercrime, bypassing security and compromising systems.
Recognizing existential threats such as these and other emerging issues with AI, the EU AI Act was proposed on 21 April 2021 by the European Commission. The Act aims to introduce a comprehensive legal and regulatory framework that encompasses all AI technologies across a broad range of sectors. It is hoped that this will protect citizens in the EU from any emerging AI-related threats.
Key Provisions and Objectives of the Act
The key objectives of the AI Act are to ensure that all AI technologies and systems used in the European Union are safe, non-discriminatory, traceable, transparent, and environmentally friendly.
The Act also says that any AI technologies should be overseen by humans to provide oversight and prevent any harmful outcomes before they occur. This task should not be left to automation, according to the Act’s authors.
Furthermore, the AI Act proposes a uniform, technology-neutral definition of AI. This could be applied to all future systems. Additionally, the Act aims to establish an EU-wide Artificial Intelligence Board to ensure compliance and promote cooperation amongst member states and industry.
Specific risk categories are proposed within the framework of the AI Act: Minimal, limited, high, and unacceptable risk. A General-purpose AI category was added in 2023 to encompass ChatGPT and other foundational models. Most systems are expected to fall into the minimal risk category, with those in the unacceptable risk category, such as systems that manipulate behavior banned outright.
The proposed regulations make exceptions for AI systems used solely for research purposes, non-professional purposes, and military and national security-related tools.
Impact of the EU AI Act on Artificial Intelligence Regulation
The EU AI Act, being the first of its kind in the world, is expected to have a profound impact on the development of AI technologies and regulation in the European Union and could provide the basis for other global economies to introduce their legal frameworks governing this emerging technological sector.
While there is a pressing need for regulatory frameworks and acts such as the AI Act to ensure that AI technologies are safe, transparent, and non-discriminatory, there is a possibility that they could stifle innovation. For this reason, the European Commission has proposed that member states establish a regulatory sandbox.
This would be a controlled environment that can be used to test, validate, and develop new, innovative AI tools before they are released on the market, ensuring that they are compliant with the Act.
Image Credit: Beautrium/Shutterstock.com
Implications for Businesses and Organizations
It is likely that the EU AI Act will have profound implications for organizations and businesses developing and using AI tools such as machine learning and Generative AI such as ChatGPT. For any current and future systems to be able to be used within the EU, they will have to comply with the new Act.
There is some uncertainty about the responsibilities and roles of different actors within the AI value chain. These are users, developers, and providers. This is highly challenging for developers of open source Generative AI that could be used by bad actors but are not designed with them in mind.
Some commentators have called for the rules to be defined according to the impact the technology has rather than the technology itself. Additionally, BEUC has proposed that consumer protection should be prioritized and guaranteed, which would require substantial improvement to the proposed Act in its current form.
Compliance costs set forth in the proposals could also have an effect on innovation in the European AI sector by businesses and organizations. AI investments could decrease by 20% over the next five years and cost the EU’s economy around €31 billion. Overly stringent conformity requirements could have a profound impact on SMEs.
Future Trends and Developments in AI Regulation
Other global economies are working on their regulatory frameworks for artificial intelligence. In the UK, a new white paper has been published, which includes £100 million in funding to help UK regulators develop systems to monitor these emerging technologies. The US has proposed an “AI Bill of Rights.”
AI technologies have the potential to vastly improve the quality of life for billions of people but come with some profound existential risks. Proposals such as the EU AI Act, whilst they would have to be tuned so as not to stifle innovation, could play a key role in ensuring that AI does not harm society.
References and Further Reading
European Parliament (2023). EU AI Act: first regulation on artificial intelligence [online] euparl.europa.eu. Available at: https://www.europarl.europa.eu/topics/en/article/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence
European Parliament (website) Artificial intelligence act [online] europarl.europa.eu. Available at: https://www.europarl.europa.eu/RegData/etudes/BRIE/2021/698792/EPRS_BRI(2021)698792_EN.pdf
Nath, P. (2024). New AI Regulation White Paper Talks of Preparing the UK Economy for Future AI Risks [online] ibtimes.com. Available at: https://www.ibtimes.co.uk/new-ai-regulation-white-paper-talks-preparing-uk-economy-future-ai-risks-1723300
Firth-Butterfield, Silverman, K & Larsen, B. (2022). Understanding the US 'AI Bill of Rights' - and how it can help keep AI Accountable [online] weforum.org. Available at: https://www.weforum.org/agenda/2022/10/understanding-the-ai-bill-of-rights-protection/
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.