EU AI Act: Regulations enfored from 2 August 2026
The EU has been the first to create regulations that govern the use of artifical intelligence.
The AI Act is the worlds first comprehensive AI law created to form part of their digital strategy.
The aim of the act is to start creating regulations in terms of development and management on the process of adopting AI agents and models to reduce risks and offering businesses confidence with using AI.
AI offers many business benefits such as cost saving in the supply chain process, healthcare and even help in harnessing sustainable projects.
The AI framework for Europe
The AI Act is about protection, monitoring, management and being able to identify issues quickly. Whilst being innovative and useful for business and Govenments, there are still a lot of unknowns and security concerns. Here are a few of the areas being addresses.
1. Transparency
From 2nd August 2026, chatbot disclosure, synthetic content marketing and deepfake labelling will be required.
2. Banned / prohibited AI
The Act bans specific AI practices such as manipulative AI, deceitful content, vulnerable people, social scoring, predictive policing i.e profiling, certain uses of biometric categorisation involving sensitive characteristics, some real-time remote biometric identification in public spaces.
3. High risk AI
High risk AI affecting business processes such as recruitment and employment, education, training, infrastructure, access to essential private/public services, access to financial services, law enforcement, migration, asylum, justice admin, biometric applications and certain safey critical products.
These High risk providers face requirements such as risk management, data quality, technical documentation, record keeping, transparency, human oversight, accuracy, quality systems, registration in an EU databse and monitoring.
Deployers also have obligations—for example, using the system according to its instructions, monitoring it, providing human oversight and addressing risks.
4. General purpose AI (genai)
Providers of General purpose AI models have the following obligations: maintaining techinical documentation, providing quality information to companies implementing AI, policy that complies with EU copyright law, detailed training content, appointing an EU rep where provider is outside the EU.
Additional requirments for GPAI models with systemic risk include: model evaluation, adversarial testing, risk assesment, serious incident reporting, cybersecurity measures.
The Commission’s GPAI Code of Practice provides a voluntary way for providers to demonstrate compliance with relevant obligations.
5. Human oversight and fundamental rights
A major theme throughout the Act is that AI shouldn’t simply make consequential decisions without appropriate safeguards.
Depending on the system, organisations need mechanisms for:
- Human supervision
- Understanding system limitations
- Monitoring performance
- Intervention or overriding the AI
- Preventing inappropriate reliance on AI
- Assessing impacts on fundamental rights.
The public sector deploying high risk systems must carry out impact assessment reports.
6. Business obligations
The act recognises Providers, Deployers, Importers, distributers, product manugactures and authorised representives.
The Act applies to companies outside of the EU delivering AI within the EU.
7. Governance and enforcement
The Act requires the European AI Office and the European AI Board, National competent authorities and market surveillance authorieties.
The EU have in place monitoring, investigating violations and imposting penalities.
The maximum fines can be substantial—up to €35 million or 7% of worldwide annual turnover for certain prohibited AI violations, with lower maximums for other categories
8. Innovation and SMEs
The AI Act also encourages AI adoption for those seeking guidance and reassurance.
Small businesses have access to testing environments and documention offering practical guideance.
Summary
This AI Act is just the begininning and will evolve as AI becomes more sophisticated but it is a good start.
We are in the early stages of adoption but it does appear there is a new role for AI Project Managers to ensure the rules are being applied.
You can request an AI Act checklist using AI!.
There will be a need for AI regulators, those checking the paperwork to ensure businesses are following the regulations.
Some may say that the EU have too many rules and regulations, but there are there to keep us all safe and sure the processes are will managed and maintained, so I hope the UK has a look at this for their own AI safeguarding policies.
As of August 2026, the UK does not have one single, comprehensive “AI Act” equivalent to the EU AI Act.
Source: Artificial Intelligence
Source: The General-Purpose AI Code of Practice | Shaping Europe’s digital future
0 Comments