Artificial Intelligence (AI) is revolutionizing our world in remarkable ways. From customized healthcare to autonomous vehicles, AI is increasingly integrated into our everyday experiences. There are only a few people nowadays who do not raise the question: what is artificial intelligence and how will it impact our daily lives? Due to the rapid advancement of these technologies, artificial intelligence law regulation has emerged as a crucial policy concern in the European Union (EU) in recent times. Policymakers have committed to crafting an Artificial Intelligence AI framework that prioritizes human well-being, aiming to ensure that Europeans can take advantage of emerging technologies that align with the core values and principles of the EU.

In April 2021, the European Commission introduced a proposal for a regulatory framework regarding artificial intelligence AI within the European Union (EU). This draft AI act marks the first- ever attempt to establish comprehensive artificial intelligence law regulation, and primarily concentrates on how AI systems are used and the potential risks associated with them. The Council reached a consensus on the general position of EU Member States in December 2021, and Parliament voted on its stance in June 2023. Currently, EU legislators are commencing negotiations to finalize this Artificial Intelligence Law.

AI act has been crafted as a comprehensive EU legislative tool that is intended to apply uniformly to all AI systems that are introduced into the market, utilized within the European Union, as well as if the users of AI systems are located in the EU. The central element of the Artificial Intelligence Law is a categorization system that assesses the potential risk an AI technology may present to an individual's well-being, safety, or fundamental rights. This framework comprises four risk levels: unacceptable, high, limited, and minimal.

Zakon o veštačkoj inteligenciji propisuje da se sistemi koji se smatraju neprihvatljivim zbog visokog nivoa rizika, kao što su ocenjivanje društvenog ponašanja od strane Vlade i sistemi za identifikaciju na javnim mestima putem biometrijskih rešenja, zabranjuju, osim u izuzetnim slučajevima. Visokorizični AI sistemi su dozvoljeni, ali programeri i korisnici su obavezni da se pridržavaju regulativa koje zahtevaju temeljno testiranje, pažljivo vođenje dokumentacije o kvalitetu podataka i odgovornost koja definiše nadzor od strane ljudi. Takođe, ovi modeli moraju biti registrovani u bazi podataka EU. Primeri AI sistema koji potpadaju u kategoriju visokog rizika, između ostalog obuhvataju autonomna vozila, pravosudne postupke, medicinske uređaje i mašine za infrastrukturu. AI sistemi koji predstavljaju „ograničeni rizik“, kao što su četbotovi, podložni su obavezama transparentnosti, što znači da korisnik mora biti obavešten da se njihova interakcija generiše pomoću veštačke inteligencije. S druge strane, Zakon o veštačkoj inteligenciji EU propisuje da se AI sistemi koji se kategorizuju kao niski ili minimalni rizik neće podvrgavati dodatnim pravnim zahtevima.

Data protection and Artificial Intelligence AI technology

Data protection is certainly one of the crucial legal topics nowadays, and the creation of legislation such as GDPR and the Law on Personal Data Protection of Serbia has changed the way companies process personal data. Consequently, businesses have increasingly prioritized investments in security to mitigate vulnerabilities and reduce the risks of surveillance, hacking, and cyberattacks.

Data protection and AI technology should be discussed together because the use of artificial intelligence often involves the collection, processing, and analysis of large amounts of data. Artificial intelligence in business implies that their operations should comply with legality, fairness, and transparency. This encompasses the obligation to offer individuals whose personal data is undergoing processing comprehensive and transparent information about the collection and handling of their data. Data controllers must also maintain a fair relationship with these individuals. Artificial intelligence companies should initiate this process by identifying the nature of the personal data being gathered, where it is stored, and who has access to it, a process known as data mapping. Subsequently, data management should adhere to legal principles. This entails creating a list of individuals whose data is being processed, clearly stating the legal grounds and purposes for data processing, and educating employees about the lawful collection and processing of personal data. Additionally, artificial intelligence companies should identify the required documents for data protection and generate them. Documenting the process also involves keeping records of individuals' requests regarding their data processing, as well as maintaining evidence of data transfers to third parties and data exports outside the country.

Intellectual Property and Artificial Intelligence AI

At the Digital Summit in Tallinn, it was highlighted that in the coming decade, as many as 50% of work tasks will be subjected to automation by the use of Artificial Intelligence (AI). AI has already undertaken some creative roles, such as writing novels, painting, generating speeches, and writing poems. Intellectual Property and AI are interdependent areas having in mind that Artificial Intelligence AI raises numerous Intellectual Property questions, such as: Who should benefit from moral and pecuniary rights on the work created by AI? Can AI be considered as an inventor and therefore enjoy copyright on the work created by AI? Can ChatGPT use my work without permission to train the system, if I have a copyright? At the moment, Case law in cases related to AI is mostly connected to the alleged copyright infringement.

According to the Serbian Law on Copyright, a work of authorship is an author's original intellectual creation, expressed in a certain form. An author is a NATURAL PERSON who has created a work of authorship, therefore AI system which is not a natural person can not be considered as an author and can not enjoy copyrights. To clarify, the person who created software is the author of that software, but he/she is not the author of photographs for example, generated by that AI system.

Verovatno će u budućnosti Zakon o autorskim pravima biti promenjen zbog tehnologija AI koje se razvijaju. Da bi se podržale investicije u AI, najverovatniji pristup autorskim pravima je onaj koji dodeljuje autorska prava subjektu koji omogućava stvaranje AI. Na primer, novi Zakon o autorskim i srodnim pravima koji uvodi AI kao stvaraoca stupio je na snagu u Ukrajini 1. januara 2023. godine. On uvodi „sui generis“ pravo koje štiti neoriginalna dela stvorena uz upotrebu AI – dela koja su stvorena bez učešća ljudi.

Liability for damage caused by Artificial Intelligence AI systems

Liability for damage caused by AI systems is a complex and evolving legal issue in today's technology-driven world. We can often hear questions, such as: is artificial intelligence dangerous? Also, as AI technology becomes more integrated into various aspects of our lives, questions arise about who should be held responsible when AI systems malfunction or cause harm. Traditional liability frameworks may not always suffice. Currently, legal frameworks are facing challenges in establishing precise rules for accountability of the AI systems. Achieving a harmonious blend between promoting innovation and ensuring liability for damage caused by AI systems is vital to creating an efficient structure that tackles the issues stemming from damages caused by AI.

One of the positive steps in that direction is certainly the EU AI Liability Directive which enhances and updates the civil liability structure of the EU, incorporating regulations tailored specifically for damages resulting from AI systems.

Legal aspects of using Artificial Intelligence AI in medicine

Utilizing AI models to analyze medical information and extract valuable insights for enhancing health results and patient well-being characterizes the application of AI in medicine. Recent progress in information technology has expedited the integration of artificial intelligence (AI) as an essential component of healthcare practices. At the moment, usually, AI is used in medicine for clinical decision support and imaging analysis. Clinical decision support systems help healthcare providers in making informed decisions about patient treatments, medications, etc. Regarding imaging analysis, it is used to examine CT scans, x-rays, MRIs, and similar.

Regarding legal aspects of using AI in medicine, it is important to point out that AI clinical decision support software gathers insights from the health data of select patients through data collection, enabling it to offer suggestions for other patients. Therefore, the doctor must handle possible concerns associated with confidentiality and privacy, as patients are entitled to understand the methods through which their confidential health data is utilized for research, diagnosis, and treatment intentions. Moreover, liability for damage caused by AI systems in medicine is also a crucial question, having in mind that AI can incorrectly diagnose or provide inappropriate treatment due to inaccuracies in AI decision-making. Assignment of liability between healthcare professionals and AI developers is a challenging task. Also, it goes without saying that most applicable healthcare laws pre-date the emergence of AI, so the laws must be amended. The use of AI in medicine certainly involves moral concerns that require our immediate attention. Ethics of AI brings up issues such as safety, transparency, algorithmic fairness, and data privacy.

Ethics of artificial intelligence AI in a legal context

Ethics constitutes a collection of ethical standards that help in distinguishing between what is morally correct and incorrect. AI ethics examines methods to maximize the positive influence of AI while minimizing potential risks and negative consequences. The performance of AI aligns with its design, development, training, and utilization. Ethics of artificial intelligence in a legal context focuses on creating a framework of ethical principles and guidelines throughout every stage of an AI system's existence.

Governments, researchers, and developers agree that an effective governance program should clarify the duties and obligations of individuals working with AI, as well as create procedures for building, supervising, and discussing AI, as well as the risks associated with it.

Case law in cases related to AI

Open AI, the company behind the well-known AI tool ChatGPT, already faces lawsuits. A class action lawsuit has been filed before the San Francisco Federal Court for supposedly breaking several privacy laws in developing the ChatGPT. It goes without saying that ChatGPT is a consumer application experiencing the most rapid growth in history, and it attracted a lot of publicity. Open AI is accused of misusing the user’s personal data from social media platforms and other sites to train ChatGPT. This is one great example of why data protection and artificial intelligence AI technology should be discussed together.

Moreover, other lawsuits against artificial intelligence companies pop up because of the alleged copyright infringement. Undoubtedly, any author shall enjoy moral and pecuniary rights concerning his/her work of authorship from the moment of its creation. The claimants – authors state that Open AI used their novels as training data for AI tool without their consent. They assert that ChatGPT produces precise summaries of multiple infringed works. Having in mind that Open AI did not license their novels, it disrupted their ability to benefit financially from their work.

When it comes to intellectual property and artificial intelligence AI, Getty Images, the stock photography company, filed a lawsuit against Stable Diffusion, text to image model, in the US for copyright infringement. Getty Images claims that the world’s leading open-source generative AI company copied more than 12 million images from its database without permission or compensation.

Case law in cases related to AI in many countries still does not exist, but is it undisputed that the outcome of these lawsuits will have an impact on it.

Artificial intelligence attorney in Serbia

The EU Artificial Intelligence Law stands as the first tangible endeavor for artificial intelligence law regulation. Its objective is to transform Europe into a prominent center for reliable AI by establishing standardized regulations that govern the creation and utilization of AI within the EU. After the terms are finalized between the European Parliament, the Council of the European Union, and the European Commission, there would be a a grace period of around two years to allow parties to comply with the regulations. After this grace period, administrative penalties of different magnitudes will be imposed - up to €30 million or 6% of the global annual turnover, serving as consequences for failure to adhere to artificial intelligence law regulation.

Based on artificial intelligence law firms’ observations, implementing new regulations consistently is frequently difficult and time-consuming. Artificial intelligence attorney in Serbia considers it vital to initiate action promptly to meet deadlines and facilitate enduring transformation and can help companies to understand and comply with the demands outlined in the EU Artificial Intelligence Law. If you need legal assistance regarding artificial intelligence law regulation, contact us by email at [email protected].

For this BLOG, the following references were used:

• European Parliament Think Tank
• EY - The EU AI Act: What it means for your business
• World Economic Forum - European Union AI Act Explained
• Reuters - OpenAI and Microsoft face new US consumer privacy class action
• Hollywood Reporter - Authors join lawsuit against OpenAI over copyright infringement
• The Guardian - Authors file a lawsuit against OpenAI for unlawfully ingesting their books
• The Verge - AI art copyright lawsuit involving Getty Images
• IBM - AI Ethics
• World Intellectual Property Organization - AI-related legal instruments