EU AI Act Compliance Introduction

Grasp the purpose of the EU AI Act and how it safeguards your business and AI users.

Explore the ethical and practical reasons driving AI regulation and how the EU AI Act balances innovation with responsibility.

Learn how biased data leads to unfair AI outcomes and ensure compliance with AI Act standards.

Discover the key factors that determine whether your systems fall under the EU AI Act regulations.

Define the key players (providers, users, etc.) and their compliance obligations under the EU AI Act.

Gain insights into the long-term impact of the EU AI Act on AI development and ethical use within Europe.

Break down the EU AI Act's structure for a clear understanding of its key components.

Know critical deadlines and understand the wide-ranging industries and individuals impacted by the Act.

Master the key vocabulary of the EU AI Act for effective communication and compliance.

Even though the thought of autonomous AI systems seems straightforward, within this video we will explore that AI system can be characterised as autonomous in various ways.

EU Guidelines are allowing us to understand why most systems will fall under the "autonomy" part of the definition.

Explore how 'learning' AI systems fall under the EU AI Act and how regulations are designed to address this.

A key characteristic of an AI system is capability to infer. Within this video, we will thoroughly explore what inference means.

Understand the EU's criteria for classifying AI, ensuring your systems align with the regulation.

Learn if your AI systems fall within the scope of EU AI Act, avoiding potential compliance issues.

EU published guidelines on the AI definition. These bring a bit more light onto the issue of whether purely rule-based systems will be classified as AI.

Term "publicly accessible space" occurs multiple times across the EU AI Act. Within this video, we will explain how EU defines publicly accessible space.

Emotion detection through AI systems is heavily regulated by the AI Act. However, within this video we will learn the crucial differences between emotions, physical states and gestures.

EU AI Act identifies two basic entities when it comes to development and use of AI systems - providers and deployers. Within this video, we will cover both of these terms thoroughly.

If you are a provider of an AI system, you need to specify for what is this model intended to be used. Moreover, if you are supposed to think about potential misuses that might arise and specify them.

Explore a key term of "profiling" that involves various activities around classification and modelling on top of pesonal data. This term will be crucial later, when we will be classifying AI systems into high-risk.

EU does not want to suppress innovation within the open-source realm when it comes to AI development. However, if a model wants to be classified as open-source by the AI Act, it has to fulfil several conditions.

Article 4 clearly outlines the obligation for companies to train their employees on AI literacy.

Within this lecture, you can find raw texts from the regulation in case you would like to get a good look at them.

Protect your business and users by steering clear of unethical practices outlawed by the EU AI Act.

Learn how the regulation combats harmful uses of AI designed to manipulate & mislead.

Discover how the EU AI Act targets AI systems that exploit the vulnerabilities of individuals or groups.

Learn about 'social scoring' systems and how the regulations ensure AI systems uphold fairness.

Understand the EU AI Act safeguards for biometric identification (facial recognition, etc.) and mass surveillance concerns.

Within this lecture, you can find raw texts from the regulation in case you would like to get a good look at them.

European Commission published Guidelines for Prohibited Practices, which will lead to the expansion of this section of the course.

Dive into the category of highest scrutiny with an emphasis on your obligations and responsibilities.

Navigate the risk-based approach with a clear understanding of how to classify your AI systems.

Even if your AI use case falls under the Annex III classification, there is a possible "way out" of this classification through possible derrogation. We will explore it in this lecture.

Within this lecture, you can find raw texts from the regulation in case you would like to get a good look at them.

Explore the stringent EU regulations on biometric systems, focusing on their role in surveillance, authentication, and how these technologies are governed under the AI Act to ensure privacy and human rights.

Understand the AI Act’s requirements for AI systems in critical infrastructure sectors, ensuring their reliability and security in vital areas like energy, transportation, and water supply.

Dive into the impact of AI on education and vocational training, examining the AI Act's guidelines for ensuring fairness, transparency, and non-discrimination in these crucial fields.

Learn how the AI Act regulates AI applications in employment, focusing on recruitment, performance monitoring, and worker management to promote fairness and transparency in the workplace.

Discover the AI Act’s provisions for AI systems that manage access to essential services and benefits, ensuring these technologies are used responsibly and equitably in public and private sectors.

Investigate the specific rules governing the use of AI in law enforcement, as outlined in the AI Act, to balance innovation with the protection of fundamental rights and freedoms.

Examine how the AI Act addresses the use of AI in managing migration, asylum, and border control, focusing on ethical considerations and the safeguarding of human rights.

Analyze the AI Act’s approach to AI systems in justice and democratic processes, ensuring transparency, fairness, and accountability in decision-making.

Stay informed on the potential future amendments to Annex III under Article 7, understanding how the AI Act can adapt to emerging risks and technological advancements.

Within this lecture, you can find raw texts from the regulation in case you would like to get a good look at them.

Create a practical roadmap for compliance with step-by-step obligations under the Act.

Within this lecture, we will examine requirements outlined in Article 9 of the EU AI Act on the establishment of risk management system for high-risk AI systems.

Requirements for data governance for high-risk AI are substantially extended by Article 10 of the EU AI Act. Let´s examine them.

Article 11 is straightforward - we are supposed to create a documentation for our high-risk AI that elaborates on all the neccessary compliance steps we took when developing and deploying it.

When we are deploying a high-risk AI system, we shall establish a thorough record keeping, or logging system. Within this lecture, we will examine simulated case study to see how this might look in practice.

If we are a provider of an AI system, selling, or otherwise providing the system to other entities (deployers), we shall provide a transparent documentation containing necessary information about this AI system.

In order for a high-risk AI system to be compliant, we need to implement functionality of human oversight, that will allow humans to step into the process automated to AI, whenever neccessary.

Article 15 brings very broad requirements on the performance (accuracy) of high-risk AI systems. Within the lecture, we will dive deeper into several examples to appreciate that in order to be compliant with Article 15, significant technical investments shall be made.

Article 15 brings very broad requirements on the robustness of high-risk AI systems. Within the lecture, we will dive deeper into several examples to appreciate that in order to be compliant with Article 15, significant technical investments shall be made.

We are finishing up our lengthy tour of Article 15 of EU AI Act. Our last topic is enhanced considerations for cybersecurity, stemmed in the use of AI systems.

Within this lecture, you can find raw texts from the regulation in case you would like to get a good look at them.

See the tangible impact on business processes, development, and deployment for high-risk AI.

Within this video, we get an overview of upcoming chapter, within which we focus on Title IV of the EU AI Act.

If a natural persons interacts with an AI system, it should be clear that this is an interaction with AI and not with another natural person.

Within this lecture, we cover a crucial coneptual difference between synthetic content and deepfakes, which are both mentioned within Title IV of the EU AI Act.

If you produce synthetic content, such as text, image or video, this shall be clearly labelled as a product of artificial intelligence (with exceptions).

Surprisingly, AI Act allows for a creation of deep fakes. This comes with a transparency obligations that we will discuss within this lecture.

Get an overview of this section that covers Title V: General Purpose AI.

Within this lecture, we take a look at how AI Act defines "general purpose AI" and thanks to that, we will learn what distinguishes these models from more casual AI models.

GPAI Q&A brought new concretizations to which models will be classified as GPAI.

If a model is classified as "general purpose AI", there is a list of specific obligations awaiting. We will cover these obligations within this lecture.

GPAI Q&A brought new concretizations to the obligations for GPAI.

Similarly to what we have seen with common AI models with the classification rules into high-risk, we now see that GPAI models can be classified into having systemic risk. In this video, we cover these classification rules.

GPAI Q&A brought new concretizations to which models will be classified as GPAI.

If a model gets classified as having "systemic risk" there is a list of obligations. We will cover these obligations, within this lecture.

GPAI Q&A brought new concretizations to the obligations for GPAI models with systemic risks.

Meet the key players enforcing the EU AI Act and their roles in ensuring compliance.

Explore the concept of 'regulatory sandboxes' and how they support AI development with compliance.

Gain your instructor's insights into the potential of sandboxes and how they can benefit businesses.

Develop a clear, step-by-step compliance plan to ensure your business aligns with EU AI Act regulations.

Understand the complex legal issues surrounding AI-generated content, who owns the rights, and how the EU AI Act might shape it.

This is a bonus lecture.