EU AI Act: What Changes for Your Website as of August 1, 2026?

The EU AI Act Becomes Reality: Is Your Website Ready for August 1, 2026?

On 1 August 2024, the EU AI Act (Regulation (EU) 2024/1689) officially entered into force, the world's first comprehensive AI legislation. However, the real impact for businesses and website owners does not begin until 2 August 2026 , when most obligations come into effect.

For Dutch and European companies using AI tools on their websites, think chatbots, content generation, or personalization, this means a fundamental shift in how you are allowed to deploy technology. AI is no longer an unregulated playing field; the European Union is now imposing strict rules that directly impact your digital presence.

In this blog post, we explain what changes as of August 1, 2026, how the current situation differs from the situation after this date, and what this specifically means for your website. We use clear comparison tables, practical examples, and a concrete action list so that you can prepare your business in time.

What is the EU AI Act and why is this law important for websites?

The EU AI Act is not just new legislation, it is a horizontal law that applies to all sectors, including websites, apps, and online platforms. Unlike sectoral legislation, the AI Act focuses on the risk that an AI system poses to users.

The Risk-Based Approach

The AI Act divides AI systems into four risk levels:

Risk level	Examples for Websites	Legal Status as of August 2026
Unacceptable Risk	Social scoring, subliminal manipulation, real-time biometric identification in public spaces	Prohibited since February 2, 2025
High Risk	Automated creditworthiness checks, CV screening, admission procedures, insurance risk analyses	Obligations as of December 2, 2027 (postponed via Digital Omnibus)
Limited Risk	Chatbots, AI content generation (blogs/images), deepfakes, emotion recognition	Transparency obligations as of 2 August 2026
Minimal Risk	Spam filters, basic optimization tools, AI games, simple recommendation engines	Exempt, voluntary codes of conduct

Important update: Due to the Digital Omnibus Agreement of 7 May 2026, the obligations for high-risk AI systems (Annex III) have been postponed from 2 August 2026 to 2 December 2027. Obligations for high-risk AI in regulated products (Annex I) apply from 2 August 2028. The transparency obligations for AI-generated content have been postponed to 2 December 2026 .

Situation Now vs. Situation After August 1, 2026: The Great Comparison

What Changes Specifically?

Aspect	Current situation (Before August 2026)	Situation AFTER August 1, 2026
AI chatbots on websites	Virtually unregulated; optional disclaimers	Mandatory transparency: users must be explicitly informed that they are communicating with AI
AI-generated content	No legal obligation to label	Mandatory labeling: AI-generated text, images, and video must be marked as synthetic (from December 2, 2026)
Product recommendations	Completely exempt; no accountability	Minimal risk: no specific AI Act obligations, but it is recommended to be transparent.
CV screening / recruitment	Only GDPR requirements apply	High risk (as of Dec 2027): human supervision mandatory, registration in EU database, fundamental rights impact assessment
Creditworthiness checks	Financial regulations + GDPR	High risk (as of Dec 2027): strict data governance, bias detection, mandatory documentation, CE marking
Deepfakes / synthetic media	Platform-dependent rules	Mandatory labeling: clear and visible indication that content is AI-generated
AI for fraud detection	Internal security protocols	Minimal risk: no specific AI Act rules, existing security standards remain applicable
Data governance for AI	Proprietary standards	Required: representative, relevant, and unbiased training data; documentation required
Human supervision	Optional or absent in many systems	Mandatory for limited and high risk: humans must be able to override and correct AI outputs.
Fines for non-compliance	No AI-specific sanctions	Up to €35 million or 7% of global turnover for prohibited practices; lower but substantial fines for other violations.

Transparency Obligations in Detail

The transparency obligations taking effect on August 2, 2026 , are the most direct and concrete consequence of the AI Act for most websites. These rules apply to low-risk AI systems, the category in which most commercial website AI falls.

AI Application on Website	Transparency requirement as of August 2026	Practical Implementation
Chatbot / virtual assistant	The user must know that they are communicating with AI.	Clear disclaimer in chat window: "You are talking to an AI assistant"
AI-generated blog posts	Content must be labeled as AI-generated	Watermark, metadata tag or visible mention above article
AI images/videos	Synthetic media must be identified	Machine-readable metadata + visible labeling for deepfakes
Emotion recognition	The user must be informed about the use	Pop-up or mention in privacy policy + real-time notification
Biometric categorization	Explicit notification required	Clear request for consent + explanation of purpose and duration

What Does the AI Act Specifically Mean for Your Website?

E-commerce Websites

For webshops, relatively little changes regarding product recommendations, these fall under minimal risk and are largely exempt. However, you should pay attention to:

• AI chatbots for customer service: As of August 2026, you must explicitly state that customers are communicating with an AI. A small line of text in the chat window suffices.

• Dynamic pricing: Adjusting prices based on supply and demand is permitted, but the use of AI to target vulnerable individuals in emergency situations is prohibited (unacceptable risk).

• Fraud detection: Background systems for credit card fraud fall under minimal risk, there are no new AI Act obligations here.

Media, Blog, and News Websites

The rules are becoming stricter for publishers and content platforms:

• AI-generated articles: If you publish news or information generated by AI without significant human oversight, this must be labeled as AI-generated content.

• Deepfakes: AI-generated images or videos that mimic real people or events must be clearly labeled to prevent deception.

• AI translations: The use of AI to translate your website is permitted and requires no special labeling, this falls under minimal risk.

Recruitment and HR Platforms

This is one of the most heavily regulated sectors under the AI Act. Although the obligations have been postponed to December 2027 , preparation is essential now:

• CV screening and candidate evaluation: AI systems that automatically filter applications or score candidates fall under high risk . You must have a human review those AI decisions and register the system in the EU database.

• Profiling: AI-based analysis and profiling can trigger both AI Act and GDPR obligations. In many cases, you need explicit consent.

Financial and Insurance Websites

• Credit scoring: AI that determines whether a visitor qualifies for a loan or credit card is high risk . You must be transparent about the decision-making and ensure unbiased data.

• Insurance risk analysis: AI for life and health insurance falls under high risk; standard car or home insurance is less strictly regulated but still subject to strict data privacy rules.

The Digital Omnibus Update: What Has Changed in May 2026?

On 7 May 2026, the Council of the EU and the European Parliament reached a provisional political agreement on the Digital Omnibus Package . This amends and simplifies certain provisions of the AI Act.

Changed Data in Overview

Provision	Original date	New date	Status
Article 5 (prohibited practices)	February 2, 2025	February 2, 2025	In operation
Article 4 (AI literacy)	February 2, 2025	February 2, 2025	In operation
GPAI obligations (Art. 51-55)	August 2, 2025	August 2, 2025	In operation
Article 50(2) watermarking & synthetic content	August 2, 2026	December 2, 2026	postponed by 4 months
National AI sandboxes	August 2, 2026	August 2, 2027	postponed by 12 months
High-risk Annex III (standalone systems)	August 2, 2026	December 2, 2027	postponed by 16 months
High-risk AI in Annex I products	August 2, 2027	August 2, 2028	postponed by 12 months

Practical consequence: As a website owner, you now have slightly more time to prepare for high-risk obligations, but the transparency rules for chatbots and AI content will still come into effect on August 2, 2026. So this is not a time for postponement, only for targeted preparation.

Compliance Checklist: How to Make Your Website AI-Act-Proof

Use this checklist to get your website ready for August 1, 2026:

Step 1: Inventory All AI Systems

Identify every AI tool running on your website, including those from third parties and vendors. Think of chatbots, content generation tools, recommendation engines, and analytics tools.

Step 2: Determine Your Risk Level

Classify each AI system according to the four risk categories of the AI Act. Don't forget to assess your third-party vendors as well.

Step 3: Implement Transparency Measures

• Add clear AI disclaimers to chatbots

• Label AI-generated content with watermarks or metadata

• Inform users about emotion recognition or biometric analysis

Step 4: Register High-Risk Systems

If you use high-risk AI systems (for example, for recruitment or credit scoring), you must register them in the EU database as soon as the obligations come into force in December 2027.

Step 5: Set up Data Governance

Ensure that your training data:

• Is representative and relevant

• Free from known bias where possible

• Is clearly documented

Step 6: Implement Human Supervision

Ensure that a human is always available to review, correct, and override AI outputs where necessary, especially for low- or high-risk systems.

Step 7: Document Everything

Keep technical documentation for:

• AI system descriptions

• Risk assessments and mitigation measures

• Training data summaries

• Human supervision practices

• Vendor compliance documentation

Step 8: Conduct Impact Assessments

For public authorities and private entities providing public services: conduct a Fundamental Rights Impact Assessment (FRIA) before implementing high-risk systems.

Step 9: Align with GDPR

Many AI Act obligations overlap with the GDPR, think of transparency, data protection, security, and consent. Ensure that your Consent Management Platform (CMP) also covers AI usage.

Step 10: Monitor Continuously

Privacy rules change. Establish internal monitoring processes to continuously comply with the AI Act, including cybersecurity, data accuracy, and robustness.

The Price of Non-Compliance: What Does It Cost If You Are Not Prepared?

The AI Act includes risk-based fines that can run up to a substantial amount:

Offence	Maximum Fine
Prohibited AI practices (Art. 5)	€35 million OR 7% of global annual revenue—whichever is higher
GPAI violations (Art. 51-55)	€15 million or 3% of global annual turnover
High-risk system requirements	€15 million or 3% of global annual turnover
Incorrect/misleading information to authorities	€7.5 million OR 1% of global annual turnover

Good news for SMEs: Lower thresholds apply to small and medium-sized enterprises (SMEs) and startups, so fines are not immediately devastating. However, even for smaller companies, a fine of several hundred thousand euros can be a heavy blow.

Why Flor-IT is Your Partner in AI Compliance

At Flor-IT, we combine in-depth technical expertise with legal insight to prepare your company for the EU AI Act. Our team of specialized consultants helps you not only with the implementation of transparency measures but also with the strategic integration of AI into your business processes.

Our Expertise:

• Legal expertise: We closely follow developments surrounding the AI Act, including the recent Digital Omnibus Agreement of May 2026.

• Technical implementation: From chatbot disclaimers to metadata watermarks, we take care of the technical realization of compliance.

• Risk assessment: We classify your AI systems and advise on mitigation strategies.

• GDPR integration: We seamlessly align your AI Act compliance with your existing privacy provisions.

Our Services for AI Act Compliance

Employ	Description	Result
AI Audit	Complete inventory and risk classification of all AI systems on your website	Clear overview of compliance status
Transparency Implementation	Technical implementation of disclaimers, labels, and metadata for AI content	Website complies with Article 50 obligations
Documentation & Governance	Drafting technical documentation, risk assessments, and supervision protocols	Ready for inspection by authorities
Training & Awareness	AI literacy training for your team (mandatory under Article 4)	Employees understand AI risks and responsibilities
Continuous Monitoring	Ongoing monitoring of legislative changes and compliance status	Always up to date with the latest regulations

Ready for August 1, 2026?

The EU AI Act is no longer a thing of the future, it is a reality that will fully take effect in two and a half months regarding transparency obligations. Although some stricter obligations have been postponed to 2027 and 2028, the time for action is now. Websites that use AI to communicate with visitors or generate content must comply with strict transparency rules as of August 2, 2026.

Contact Flor-IT today for a no-obligation AI scan of your website. Together, we will ensure that you are not only compliant, but that your AI deployment becomes a competitive advantage, instead of a legal risk.

📧 info@flor-it.com | 🌐 www.flor-it.com | 📞 Contact us for a no-obligation consultation

The Future of AI Is Regulated—And That Is Good News

The EU AI Act marks the end of the "Wild West" era of artificial intelligence. For website owners, this means more responsibility, but also more user trust. By being transparent about AI usage, handling data carefully, and ensuring human oversight, you not only comply with the law, you also build a brand that consumers trust.

The main takeaways:

• August 2, 2026: Transparency obligations for low-risk AI (chatbots, content generation) take effect

• December 2, 2026: Watermarking and labeling of synthetic content mandatory

• December 2, 2027: High-risk obligations for standalone systems

• August 2, 2028: High-risk AI in regulated products

Start your preparation today. The costs of non-compliance are astronomical, but the investment in compliance is manageable, especially with the right partner by your side.

Sources:

• European Commission - AI Act Timeline & Risk-based Approach

• CookieScript - EU AI Act Checklist for Websites

• VerifyWise - EU AI Act Omnibus Changes (May 2026)

• Legal Nodes - EU AI Act 2026 Updates

• Kennedys Law - EU AI Act Implementation Timeline