The EU AI Act is the first comprehensive legal framework regulating artificial intelligence. It became EU law in August 2024. Most of its provisions take effect in stages through 2026, with the broadest requirements applying from August 2026 onwards.
Project managers do not need to become compliance specialists. But if your projects involve AI tools -- and almost all of them do now -- you need to understand what the Act requires and what your role is.
What the EU AI Act Actually Does
The Act does not ban AI. It classifies AI systems by risk level and assigns different obligations to each tier.
Unacceptable risk: AI systems that are banned outright. Examples include social scoring systems and certain biometric surveillance applications. Most project management tools do not fall here.
High risk: AI systems that must meet strict requirements before deployment. This includes AI used in critical infrastructure, hiring and HR decisions, credit scoring, law enforcement, and certain safety-critical applications. High-risk AI requires documentation, human oversight, transparency, and registration in an EU database.
Limited risk: AI systems with transparency obligations. Chatbots, for example, must tell users they are interacting with AI.
Minimal risk: The majority of AI tools -- including most productivity tools project managers use -- fall here. These have no specific obligations beyond standard good practice.
What This Means for Your Project
Your immediate responsibility as a project manager is knowing which tier each AI tool your project uses falls into.
For most project teams, this means auditing the tools currently in use and asking a few questions:
- Is this tool used to make decisions that significantly affect people? (HR systems, resource allocation with employment consequences, patient-facing healthcare tools)
- Does the vendor claim EU AI Act compliance? Do they have a conformity declaration?
- If the tool touches regulated data or regulated processes, has legal reviewed its use?
The EU AI Act holds both AI system providers and deployers accountable. As a deployer -- the organisation using the tool in a project -- you have obligations that include human oversight of high-risk AI decisions, record-keeping, and user transparency.
What High-Risk AI Requires
If any AI tool in your project qualifies as high-risk under the Act, you need to:
- Document the system's purpose, capabilities, and limitations
- Establish a human oversight mechanism (someone accountable for reviewing AI outputs before consequential decisions are made)
- Maintain logs sufficient to demonstrate compliance if audited
- Ensure users understand when AI is making or influencing decisions that affect them
For most projects, this is not a heavy lift. The harder problem is identifying which tools are in scope -- which requires the kind of structured evaluation the Act effectively mandates.
The August 2026 Deadline
The broadest AI Act provisions take effect in August 2026. That sounds distant, but compliance readiness requires time: vendor contracts may need addenda, HR and legal need to be briefed, and high-risk AI systems may need configuration changes.
Project managers who start the inventory and classification process now will have six to twelve months to resolve compliance gaps without a deadline crisis.
The practical first step is an AI tool inventory. List every AI tool your team uses -- directly and embedded in other software (Copilot in Office 365, AI features in your project management platform, transcription tools). Classify each by the risk tier framework above.
From that list, the high-risk and limited-risk items surface. Those are the ones that need attention before August 2026.
Free tool: The AI Tool Risk Evaluator walks your team through a structured 5-minute risk evaluation before adopting any AI tool. Free, no login required.
Subscribe to AI for Project Managers at Indigo.pub for practical AI governance guidance delivered to your inbox.