What is the AI Act?
The AI Act is a European Union regulation designed to regulate the use of artificial intelligence (AI). It classifies AI systems according to their level of risk, and imposes proportionate obligations for each category.
The purpose is to guarantee security, transparency and respect for users’ fundamental rights.
In addition, the aim is to boost the competitiveness and innovation of companies in the field of AI.
As a reminder, unlike EU directives, the regulations do not require passage through the law of each country. They therefore apply directly.
It is important to bear in mind that generative AI is just one of many aspects of AI, and therefore of the corresponding Law.
The drafting process for the AI Act was already underway before the explosion of generative AI. It therefore invited itself to the Act’s table along the way.
What is the timeline for implementation of the AI Act?
The AI Act was passed in May 2024. It will come into force 20 days after its publication in the Official Journal of the European Union on July 12.
The Regulation will be applicable from August 2, 2026, with the following exceptions:
Dispositions concernées
Compliance deadlines
- Prohibitions on unacceptable risk AI
February 2, 2025
- Obligations for providers of general purpose AI models
- Appointment of Member State competent authorities
August 2, 2025
- Obligations on high-risk AI systems specifically listed in Annex III
August 2, 2027
A wealth of literature on the AI Act
Here are a few (almost arbitrary) references to help you get started:
European Commission (reference texts)
- The EU Artificial Intelligence Act
- High-level summary of the AI Act
- EU AI Act Compliance Checker
- AI Pact
Autorité de la concurrence
CNIL
Gide Loyrette Nouel
Wavestone
- AI Act: Understanding and Implementing European Law on Artificial Intelligence
- AI Act All you need to know to understand and comply with the EU law on AI
Numeum
Mlex Shorthandstories
Des « think tanks » :
- La Villa Numéris (in French): Revue sur l’IA et IA et régulation, un éternel débat…
- The Robert Schuman Foundation : What to take away from the European law on artificial intelligence
Trade and business press (in French):
What is AI (for the Law)?
The AI Act defines an AI system as follows:
« AI system means a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments ».
A wide scope
The scope of the AI Act is very broad:
- providers who market or commission AI systems or market AI models for general use in the Union, whether these providers are established or located in the Union or in a third country;
- deployers of AI systems who are established or located in the Union;
- providers and deployers of AI systems who are established or located in a third country, when the results produced by the AI system are used in the Union;
- importers and distributors of AI systems;
- product manufacturers who place an AI system on the market or in service with their product and under their own name or trademark;
- authorized representatives of service providers not established in the EU;
- the people concerned in the Union.
AI Act bends over backwards to define risk
AI Act classifies situations according to the risks to users:
- Unacceptable, prohibited risks (e.g. social rating systems and manipulative AI);
- High-risk AI systems, which are the focus of most of the text;
- Limited-risk AI systems, subject to lighter obligations (deepfakes chatbots…)
- Minimal-risk systems, which are not regulated (video games, certain anti-spam filters, etc.).
Impact on complex value chains
There are many situations with a long value chain.
It is important to ensure that the chain of shared responsibility is clear.
While the conversational AI provider is responsible for the integration and use of the LLM, the company remains accountable to the end-users.
To sum up, it is crucial to ensure that suppliers all along the chain comply with the relevant regulations.
Business applications (ERP, CRM, HRIS, etc.)
Embedding AI in your application and contributing to productivity
An increasing number of business applications have already integrated AI, or are planning to do so in the near future. This is highly relevant, given that these applications contribute to business productivity.
And integrating functionalities supported by AI technologies, in particular conversational or generative AI, boosts this productivity tenfold:
- Chatbot for querying product or business knowledge bases;
- Transactions conducted in natural language with the application, complementing traditional web interfaces or mobile applications;
- Generation of summaries, sales proposals, marketing content, etc…
What does the AI Act mean for publishers who integrate AI into their products?
So software publishers have a choice to make.
- They can either develop all or part of these functionalities themselves,
- or integrate specialized solutions that provide these capabilities with a publisher-class level of performance and security.
With rare exceptions, such business software will be considered as limited-risk AI systems.
They essentially have transparency requirements towards end users:
- AI systems intended to interact directly with natural persons shall be designed and developed in such a way that the natural persons concerned are informed that they are interacting with an AI system, unless this is obvious from the point of view of a reasonably well-informed, observant and knowledgeable natural person, taking into account the circumstances and context of use.
- Providers of AI systems that generate synthetic audio, image, video or text content ensure that the results of the AI system are marked in a machine-readable format and detectable as artificially generated or manipulated.
- Emotional recognition or biometric categorization systems must inform the natural persons who are exposed to them of how the system works and handle personal data appropriately.
When should you panic?
It is highly likely that these texts will continue to evolve, with the proliferation of technologies and the multiplication of uses.
The AI Act establishes a hybrid governance framework, with implementation and enforcement powers divided between the Union and national levels:
- The European Commission will remain central to the governance and implementation of the AI Act.
- It will be responsible for enforcing the provisions relating to GPAI models. But it will also be responsible for harmonizing the application of the AI Act across the Union. It will also define compliance with the AI Act and update certain key aspects of the regulations.
- At national level, regulators will be responsible for enforcing provisions relating to prohibited and high-risk AI practices.
It seems to us that the first step is to familiarize ourselves (a little) with this theme. Then, determine the situation in which you find yourself. And finally, to deduce the measures to be taken to ensure compliance when the law comes into force.
AI Act: to be continued...
Under those circumstances, it’s a safe bet that, like any complex and uncertain system, a running-in phase will be necessary. And this will be just as true for the legislator as it will be for those subject to the law. Texts will have to be adapted and practices put in place in line with the “common sense”. The GDPR has already given us an example on the scale of how to proceed..
_____________
Are you a software publisher? We hope this article has given you a better understanding of the AI Act.
Agora Software is the leading provider of conversational AI solutions for software publishers. We rapidly deploy conversational application interfaces that enhance the user experience of applications and platforms. We create rich, multilingual and omnichannel interactions with all your users.
Planning to integrate conversational AI into your applications?
Let’s talk: contact@agora.software
If you enjoyed this article, you might like to read 8 questions to address when integrating AI into your application.
Join us on our Linkedin page to follow our news!
Want to understand how our conversational AI platform
optimizes your users’ productivity and engagement by effectively complementing your business applications?