As we approach 2025, the European Union’s landmark Artificial Intelligence Act (EU AI Act) is set to reshape the AI landscape for companies operating in the EU. Designed to harmonize AI regulations, the Act is focused on promoting innovation while safeguarding fundamental rights, privacy, and transparency. Here's a breakdown of what’s coming and how businesses can prepare for this legislation next year.

What’s Happening in 2025?

The EU AI Act sets a timeline (see attached) for compliance that begins with foundational governance and moves toward full enforcement over several years. In 2025, several key provisions will come into effect:

February 2, 2025: AI Literacy and Training Obligations

• Companies must implement training and awareness programs for employees and other stakeholders interacting with AI systems.
• These obligations aim to foster informed usage and operation of AI systems.

 February 2, 2025: Prohibited Practices Enforcement

• Practices like AI-based social scoring, manipulative subliminal techniques, or systems exploiting vulnerabilities will be prohibited.
• Businesses need to audit and eliminate such practices to comply with Chapter II of the Act.

Use our EU AI Act checker to determine if your system will be prohibited

 May 2, 2025: General-Purpose AI Model Codes of Practice

• The European AI Office is tasked with publishing best practices for general-purpose AI models (GPAIMs). If these codes are not finalized by August 2025, the European Commission will step in to provide common rules.

 August 2, 2025: Penalties and Obligations Begin

• Enforcement of penalties begins for breaches, alongside obligations for GPAIMs released after this date.

Industries and Companies Most Affected

The Act primarily impacts companies across these domains:

High-Risk AI Systems

Sectors like healthcare, finance, and transportation will need to meet stringent safety, transparency, and accountability standards (Chapter III).

General-Purpose AI Providers

Providers of foundational AI models used across multiple applications, such as large language models or image recognition frameworks. As outlined in Chapter V, such providers must follow codes of practice to ensure their models are safe, fair, and robust. These rules aim to prevent misuse while facilitating integration into diverse industries.

Consumer-Facing AI

Companies using AI to interact with individuals or generate content (e.g., chatbots or recommendation systems) must ensure compliance with transparency requirements. Companies interacting directly with consumers using AI must comply with transparency obligations in Chapter IV. They are required to disclose AI involvement, offer understandable explanations of outputs, and protect users from deceptive practices.

Even smaller organizations and startups will need to comply, as the Act applies to all businesses offering AI systems in the EU, regardless of their geographic base. However, there are regulatory sandboxes that can help test and develop AI solutions under the supervision of regulatory authorities without facing immediate compliance penalties.

‍

Related Article: Transforming Blockchain Security: AuditOne’s AI Audit Agents

Steps Companies Should Take in 2025

To stay compliant and competitive in 2025, businesses should act now:
‍

• Evaluate existing AI systems against the prohibited practices outlined in Chapter II.

• Identify systems that might qualify as "high-risk" or general-purpose under the Act.

• Educate staff on AI ethics, safety, and compliance to align with the training mandates.

• Establish clear accountability structures for AI oversight, including compliance officers and robust documentation processes.

• Ensure systems interacting with users can disclose AI involvement and provide understandable explanations for decisions or outputs.

Looking Ahead

While 2025 is a pivotal year, the Act’s full implementation extends into the late 2020s. High-risk AI obligations, for instance, will see additional requirements phased in through 2026 and beyond​​.

For companies, this period represents both a challenge and an opportunity. By adopting compliant practices early, businesses can foster trust, reduce regulatory risks, and position themselves as leaders in responsible AI innovation. AuditOne is here to help. We provide EU AI Act compliance audits, consultations, and monitoring. Reach out to us at hello@auditone.io

‍