SynavistraTogether we chart a path forward.

EU AI Act: Risk-Based Regulatory Framework for Artificial Intelligence

Enforcement Timeline [gn400]
Governance and Enforcement [gn900]
Cross-Cutting Obligations [gn800]
Limited Risk Obligations [gn550]
General-Purpose AI Models [gn600]
Obligations by Role [gn700]
requires disclosure by [kr024]
extends to [kr025]
triggers additional obligations under [kr028]
go beyond [kr029]
grants presumption of conformity with [kr030]
cooperate with [kr032]
carry out [kr033]
becomes a provider [kr034]
requires [kr006]
requires [kr006]
takes into account [kr037]
enabling affected persons to [kr039]
supports harmonisation of [kr042]
followed by [kr020]
followed by [kr048]
followed by [kr021]
followed by [kr022]
regulates AI systems on the EU market [kr009]
applies to [kr004]
followed by [kr008]
enables [kr010]
unlimited jurisdiction to review [kr045]
in effect from [kr013]
falls under [kr005]
requires [kr006]
requires [kr006]
falls under [kr005]
requires [kr006]
in effect from [kr013]
are exempt from [kr046]
voluntarily apply [kr038]
enforce compliance for [kr041]
carry out enforcement at [kr043]
covers the majority of [kr044]
prohibits [kr007]
classified above [kr001]
must comply with [kr002]
falls under [kr005]
requires [kr006]
risk level above [kr011]
must comply with [kr012]
in effect from [kr013]
full obligations from [kr014]
full enforcement from [kr015]
apply to [kr023]
apply separately from [kr026]
underpin [kr027]
acts as [kr031]
enforced from [kr003]
apply across all risk tiers [kr036]
oversees [kr040]
governance mechanisms for [kr035]
risk level above [kr011]
enforced from [kr003]
penalises violations of [kr047]
penalises violations of [kr047]
must comply with [kr002]
must comply with [kr002]
apply across all risk tiers [kr036]
2 February 2025 [ke004]
2 May 2025 [ke005]
2 August 2025 [ke001]
2 August 2026 [ke002]
2 August 2027 [ke003]
AI Regulatory Sandbox [ke010]
EU AI Office [ke031]
European AI Board [ke033]
National Competent Authorities [ke044]
Penalty: Prohibited Practices [ke050]
Penalty: Operator Obligations [ke049]
Penalty: Incorrect Information [ke048]
GPAI Model Fines [ke036]
EU Database [ke032]
Advisory Forum [ke008]
AI Literacy [ke009]
GDPR and Data Protection [ke035]
Non-Discrimination [ke045]
Voluntary Codes of Conduct [ke071]
Right to Explanation [ke059]
Whistleblower Protection [ke073]
Right to Complaint [ke058]
Compliance Requirements [gn300]
informs [kr016]
documented in [kr017]
enables [kr018]
validated by [kr019]
Risk Management System [ke060]
Data and Data Governance [ke020]
Technical Documentation [ke067]
Transparency [ke068]
Human Oversight [ke040]
Accuracy, Robustness and Cybersecurity [ke006]
Overall Compliance [ke047]
Record-Keeping [ke057]
Provider Obligations [ke055]
Deployer Obligations [ke023]
Importer Obligations [ke041]
Distributor Obligations [ke024]
Value Chain Responsibilities [ke070]
Fundamental Rights Impact Assessment [ke034]
Post-Market Monitoring [ke052]
Serious Incident Reporting [ke061]
CE Marking [ke015]
Authorised Representative [ke012]
Systemic Risk Classification [ke065]
GPAI Technical Documentation [ke037]
Downstream Provider Information [ke025]
Copyright Compliance [ke018]
Systemic Risk Obligations [ke066]
Codes of Practice [ke016]
Open-Source Exception [ke046]
Human Interaction Disclosure [ke039]
Synthetic Content Marking [ke064]
Emotion/Biometric Disclosure [ke027]
Deep Fake Disclosure [ke021]
High-Risk Classification [gn200]
Biometrics [ke014]
Critical Infrastructure [ke019]
Education and Vocational Training [ke026]
Employment and Worker Management [ke028]
Essential Services [ke029]
Law Enforcement [ke042]
Migration, Asylum and Border Control [ke043]
Administration of Justice and Democracy [ke007]
High-Risk Derogation [ke038]
Prohibited Practices [gn100]
Subliminal Manipulation [ke063]
Vulnerability Exploitation [ke072]
Social Scoring [ke062]
Predictive Policing [ke053]
Untargeted Facial Scraping [ke069]
Workplace/Education Emotion Inference [ke074]
Biometric Categorization [ke013]
Real-Time Remote Biometric ID [ke056]
EU AI Act [ke030]
Artificial Intelligence [ke011]
Provider [ke054]
Deployer [ke022]
Placing on the market [ke051]
Conformity assessment [ke017]