As artificial intelligence reshapes our world in 2025, a wave of groundbreaking legislation is taking effect across major jurisdictions. From the EU's comprehensive AI Act to California's suite of AI-focused bills addressing everything from digital replicas to healthcare utilization reviews, governments are racing to establish guardrails for this transformative technology. This landmark year marks the first time we're seeing coordinated, large-scale regulation of AI systems, with particular attention being paid to protecting individual rights, ensuring algorithmic fairness, and maintaining human oversight in critical decisions.
Europe
EU Artificial Intelligence Act
Key Provisions Effective in 2025
The EU Artificial Intelligence Act establishes a regulatory framework to ensure the safe and ethical development of AI across Europe. Specific provisions begin applying in 2025 to address risks, ensure accountability, and set compliance timelines for various stakeholders.
Effective Dates and Provisions:
February 2, 2025:
Prohibitions on certain AI systems, as outlined in Chapter 1 and Chapter 2, come into force, marking the first application of the Act's requirements.
May 2, 2025:
The European Commission must finalize codes of practice for AI providers and other stakeholders to guide implementation.
August 2, 2025:
Application of rules regarding:
Notified bodies (Chapter III, Section 4)
General-purpose AI models (GPAI; Chapter V)
Governance and confidentiality provisions (Chapter VII and Article 78)
Penalties and enforcement mechanisms (Articles 99 and 100)
Providers of GPAI models in use before this date must achieve compliance with the Act by August 2, 2027.
Member States must designate competent authorities for market surveillance and penalties, communicating these to the Commission.
Annual Commission review begins to assess prohibitions and update guidelines as necessary.
August 2, 2025 (and every two years):
Member States must report on the financial and human resources of national authorities overseeing AI.
United States
California
AB-2602 Contracts against public policy: personal or professional services: digital replicas.
Effective Date: January 1, 2025
California's Assembly Bill 2602 introduces new provisions to protect individuals from unfair agreements involving the use of digital replicas for personal or professional services. Key provisions include:
Unenforceable Agreements:
Any contract allowing the use of a digital replica of an individual’s voice or likeness in place of the individual’s personal work is unenforceable unless:
It includes a specific description of the intended uses of the digital replica.
The individual was represented during negotiations by legal counsel or a labor union.
Definition of Digital Replica:
A "digital replica" is defined as a computer-generated, highly realistic representation of an individual’s voice or likeness in sound recordings, images, audiovisual works, or transmissions.
Excludes remastering or remixing of authorized sound or audiovisual works.
Exceptions:
Provisions with intended uses consistent with the original contract terms are exempt from unenforceability.
Exclusivity clauses in contracts are unaffected by this laws
Keep reading with a 7-day free trial
Subscribe to The AI Policy Newsletter to keep reading this post and get 7 days of free access to the full post archives.