EUROPEAN UNION ARTIFICIAL INTELLIGENCE LAW PROPOSAL (Artificial Intelligence Act)
I.Entry
Artificial intelligence has the potential to provide various economic and social advantages in many sectors and social activities and is developing rapidly. In this context, on 21.04.2021, the Commission of the European Union (“Commission”) “Establishing Harmonized Rules and Amending Certain European Union Regulations on Artificial Intelligence” (the “Proposal”) the regulation proposal named “Artificial Intelligence Regulation”. With this proposal, it is a legal regulation that includes regulations on artificial intelligence.
The proposal submitted by the Commission requires the approval of the European Parliament and member states to enter into force. If the relevant regulation becomes law, the activities of technology companies operating in other countries operating in EU countries will also have to comply with the relevant law.
The ultimate aim of the EU Commission’s proposal is to ensure a controlled progression of AI systems, with the ultimate aim of preventing harmful uses of AI tools and eliminating dangers such as misdirection and limiting future risks.
Finally, the proposal aims to foster trust by establishing a legal framework. Moreover, the proposal is based on EU values and fundamental rights and freedoms and aims to encourage the adoption of AI-based solutions.
II.What is the Scope of the Law Proposed by the Commission?
The current regulation is expected to affect providers placing AI systems on the market or making them available in the EU, distributors of AI systems operating in the Union, importers of AI systems where they meet certain criteria, and distributors and manufacturers.
With the aforementioned regulation, if the outputs of artificial intelligence systems produced within the EU are used, artificial intelligence system providers operating outside the EU borders will also be subject to the relevant regulation. This approach is likely to have a global impact on stakeholders in the technology sector around the world.
III.Some Issues Regulated by the Proposal Submitted by the EU Commission
The European Parliament’s priority is to ensure that Artificial Intelligence (“AI“) systems used in the EU are safe, transparent, traceable, non-discriminatory and environmentally friendly. AI systems should be supervised by humans rather than automation to prevent harmful consequences.
Furthermore, the European Parliament wants to establish a technology-neutral, uniform definition for AI that can be applied to future AI systems. Under the proposal, artificial intelligence is defined as follows: ““Artificial intelligence system” (AI system) means software developed with one or more of the techniques and approaches listed in Annex I and capable of producing outputs such as content, predictions, recommendations or decisions that affect the environments in which they interact, for a given set of human-defined objectives.”
The European Parliament has defined artificial intelligence systems under four main headings as “Low Risk, Limited Risk, High Risk and Unacceptable“.
-
Under Article 5 of the proposed AI law, if it poses a clear threat to people’s safety, livelihoods and fundamental rights due to an “Unacceptable Risk“, its placing on the market, making available or use would be prohibited in the EU and, if approved, in Member States.
-
Article 6 of the proposed AI law regulates “High Risk” AI systems that have an adverse impact.
According to Art.6/2 of the AI law, high-risk AI systems are AI systems listed in any of the following areas:
- Biometric Identification and Classification of Natural Persons:
- Infrastructure Management and Operation:
- Education and Vocational Training:
- Employment, Labor Management and Access to Self-Employment:
- Access to and utilization of essential private and public services and assistance:
- Law Enforcement:
- Migration, Asylum and Border Control Management:
- Administration of Justice and Democratic Processes:
In this context, providers, importers, distributors and users of high-risk AI systems will have to fulfill a number of obligations.
- Providers of high-risk AI systems will first need to register the model system in an EU-wide database managed by the Commission before placing their system on the market or putting it into service.
- Risk management should be established.
- The AI system should be developed with transparency, openness and auditability.
- Conformity assessments of AI systems should be carried out before they are placed on the market.
- Once the system is on the market, systems for surveillance and monitoring of people should be developed. In order to prevent risks to fundamental rights and freedoms, the system should be designed in such a way that the relevant users are informed.
- For a high-risk AI system, it must undergo a compliance assessment in any case. (Compliance Assessment)
It is recommended that the ( Code of Conduct ) on artificial intelligence should apply to all AI systems, regardless of risk group.
-
If the proposed AI system is “Limited Risk” or “Low Risk”, it is deemed sufficient for the AI system to provide information only.
The table below refers to the risk groups specified in the Proposal.
| Unacceptable Risk | Cognitive behavioral manipulation of people or certain vulnerable groups, voice-activated toys that encourage dangerous behavior in children, and real-time and remote biometric identification systems such as facial recognition |
|
High Risk
|
Surgical robots with Autonomous Devices |
| Limited Risk |
AI systems that present “limited risk” such as systems that interact with humans (e.g. chatbots) |
| Low and Minimal Risk | All other low and minimal risk AI systems can be developed and used in the EU without any legal obligation |
IV.What are the Sanctions Under the Proposal?
As a result of violating the obligations and rules set out under this proposal, leading companies in the technology sector may suffer great material and moral damage. This is because companies such as OpenAI, ChatGPT, Google Bard and Midjourney, which develop artificial intelligence models at the production stage, may be subject to severe sanctions as a result of violating the specified obligations. This will require a more secure and transparent standard to be set across the globe.
Enforcement of the rules and obligations will be up to the 27 member states of the EU and regulators will have the power to force non-compliant companies to withdraw their apps from the market. The relevant AI law provides for fines for breaches of obligations and rules, depending on the severity of the breach, ranging from €10M to €40M or up to 7% of the company’s annual global turnover. This could lead to significant losses, especially for tech giants such as Google and Microsoft.
V.Conclusion
In light of all these explanations, the main purpose of the proposed law is to monitor the potential risks of artificial intelligence technology, prevent the potential for abuse and ensure the progress of technology in a controlled manner. With this regulation, it has the potential to affect technology companies using today’s algorithms and artificial intelligence. Ultimately, in the following process, member states are expected to come together to determine the final text.
Bibliography
EU Artificial Intelligence Law: First regulation on artificial intelligence
“EU Legislation in Force” Briefings
Artificial Intelligence Law: One step closer to the first rules on Artificial Intelligence
Regulation of the European Parliament and of the Council laying down harmonized rules on artificial intelligence (Artificial Intelligence Act) and amending certain Union laws