- with readers working within the Banking & Credit, Business & Consumer Services and Media & Information industries
- within Strategy topic(s)
Legal Developments in AI Applications: A Brief Overview of 2025
As artificial intelligence ("AI") technologies continue to permeate economic, social, and institutional structures at an increasingly profound level, legal systems across the globe are swiftly developing regulatory frameworks to keep pace with this transformation. As of 2025, numerous jurisdictions have begun addressing issues of liability, accountability, and human rights in the context of AI deployment.
European Union: The AI Act as a Global Benchmark
The EU Artificial Intelligence Act ("Act"), entering into force on 1 August 2024, remains the first comprehensive and binding AI-specific legal framework. Adopting a risk-based approach, it classifies AI systems into "unacceptable," "high," "limited," and "minimal" risk categories, imposing obligations proportionate to the risks involved.
Unacceptable-risk systems – including those employing manipulative techniques, social scoring, biometric categorization, or real-time biometric surveillance for law enforcement – are strictly prohibited.
High-risk systems, covering areas such as employment, education, critical infrastructure, and electoral influence, must comply with stringent governance standards. These include documented risk management processes, high-quality datasets, traceability, and logging obligations, human oversight mechanisms, and demonstrable levels of accuracy, robustness, and cybersecurity. Providers must maintain detailed technical documentation, establish a risk management system, and ensure transparency in AI-human interactions.
The Act also addresses general-purpose AI ("GPAI") models, which underpin numerous AI applications. These models are subject to transparency and intellectual property safeguards, with compliance obligations taking effect in August 2025 and full implementation due by August 2027.
The sanctions set forth under the Regulation are notably severe: the use of prohibited systems may result in an administrative fine of up to EUR 35 million or 7% of the global annual turnover, whichever is higher. Other violations are subject to fines up to EUR 15 million or 3% of the turnover, again whichever is higher. The implementation and supervisory processes are carried out by the European AI Office in cooperation with the competent authorities of the Member States. In mid-2025, the European Parliament published a detailed implementation timeline.
The Act establishes a harmonized framework that sees to balance innovation and protection, offering legal certainty to AI developers and operators while reinforcing the EU's commitment to human rights and data security.
United States: Fragmented Yet Fast-Moving
Although a federal-level AI law has not yet entered into force in the United States, numerous state-level regulations have rapidly emerged, focusing on consumer protection and the safeguarding of individual rights.
- Tennessee ELVIS Act (2024) prohibits the unauthorized use of an individual's voice, likeness, or image – an early response to generative AI's capacity for deepfakes.
- Utah's AI Amendments Act (Senate Bill 149) mandates disclosure when generative AI systems are used to produce text, visual, or audio outputs, particularly in healthcare contexts. Utah remains the first state to regulate private-sector AI transparency comprehensively.
- Colorado Artificial Intelligence Act (CAIA), scheduled to enter into force in 2026, imposes duties of care and anti-discrimination obligations on developers and deployers of high-risk AI systems in sectors such as finance, housing, employment, and healthcare.
At the federal level, the "Take It Down Act" marks a significant milestone. It criminalizes the creation and online distribution of AI-generated sexually explicit content produced without consent, mandating that covered platforms remove such material within 48 hours of notification. The law is enforceable under the Federal Trade Commission Act (FTCA) and reflects growing legislative concern over AI-related privacy violations, particularly affecting minors.
In September 2025, the Federal Trade Commission (FTC) launched an inquiry into major consumer-facing AI chatbot providers, seeking to determine whether adequate safeguards exist to protect minors and assess potential adverse impacts of AI interactions on children and adolescents.
These fragmented yet proactive measures signal an evolving regulatory landscape in the United States, where states continue to lead in defining AI ethics and liability frameworks pending comprehensive federal legislation.
United Kingdom: Balancing Innovation and Copyright
The United Kingdom has yet to adopt binding AI-specific legislation. Its current pro-innovation regulatory approach emphasizes flexibility, encouraging AI development within existing legal structures. However, the absence of binding rules leaves individuals potentially exposed to rights infringements and technological misuse.
Ongoing debates indicate that the proposals concerning artificial intelligence have been postponed due to the preparation of a comprehensive bill aimed at regulating both the technology itself and the use of copyright-protected materials.
Australia: Ethical Guidance and Voluntary Standards
Australia has pursued a principles-based and voluntary approach to AI regulation. The 2019 Artificial Intelligence Ethics Principles, derived from the OECD AI Principles, outline eight voluntary guidelines promoting fairness, transparency, and accountability in AI systems.
Complementing these principles, the Voluntary AI Safety Standard provides practical guidance to organizations on mitigating risks while harnessing the benefits of artificial intelligence. It establishes ten core safeguards addressing transparency, accountability, and risk management.
Although not legally binding, these instruments can be said to establish a coherent ethical framework in Australia that may, in time, evolve into binding regulatory measures.
South Korea: The Second Comprehensive AI Law
The Act on the Development of Artificial Intelligence and the Establishment of Trust, scheduled to enter into force on 22 January 2026, is expected to become the world's second comprehensive AI legislation following the European Union's AI Regulation.
Its objectives include protecting human rights and dignity, enhancing quality of life, and strengthening national competitiveness. The legislation envisions a trusted ecosystem for AI development by defining obligations for both public and private entities and setting foundational principles for safe and ethical AI use.
Türkiye: Laying the Legislative Groundwork
As of September 2025, Türkiye has not yet enacted binding AI-specific legislation. However, significant progress has been made through strategic initiatives.
The National Artificial Intelligence Strategy (2021-2025), jointly prepared by the Presidential Digital Transformation Office and the Ministry of Industry and Technology, has been finalized and publicly released.
The subsequent 2024-2025 Action Plan focuses on cultivating artificial intelligence experts and enhancing employment in the field, fostering research and entrepreneurship, facilitating adaptation to socioeconomic transformation driven by AI, and improving access to high-quality data and infrastructure.
In October 2024, the Grand National Assembly of Türkiye ("TBMM") resolved to establish a Parliamentary Research Commission with the mandate to examine the societal impacts of artificial intelligence and identify areas requiring legislative reforms.
The Draft Artificial Intelligence Act, submitted to the Grand National Assembly of Türkiye in June 2024, remains under review by the relevant parliamentary committees. The bill aims to ensure the ethnical and safe use of artificial intelligence, safeguard privacy, and establish a comprehensive legal framework for its application across various domains.
Global Trends: Convergence Toward Shared Principles
Across more than 60 jurisdictions, AI policy discussions have intensified. Despite differing legal approaches and institutional capacities, most frameworks converge around the following three key themes:
- Managing risks associated with high-impact applications (healthcare, finance, employment, and deepfakes);
- Enhancing algorithmic transparency and accountability; and
- Ensuring data protection and human oversight.
Conclusion: Regulation Enters a Binding Phase
As of 2025, it is observed that artificial intelligence regulations have evolved from ethical declarations into binding legal norms. The EU has set the global standard through its comprehensive, risk-based model; the U.S. continues to evolve through a patchwork of state initiatives; the UK, Australia, South Korea, and Türkiye are each developing frameworks reflecting their institutional priorities.
Despite divergent paths, these jurisdictions share a common objective: ensuring that AI serves humanity safely, transparently, and accountably.
For legal professionals, understanding these fast-moving developments is essential -not only to support corporate compliance but also to safeguard fundamental rights such as privacy, data protection, and freedom of expression in the age of artificial intelligence.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
