Why the EU AI Act matters for deepfakes, fraud and business confidence

Have you ever wondered if you are dealing with a human or an AI-generated system?

The law has moved quickly in the EU. Since 2 August 2026, certain AI systems must be designed so that people can understand when they are interacting with AI or viewing AI-generated content.

What does this mean for businesses?

2 August 2026 brought Article 50 into being. It is a broad transparency regime covering the disclosure of AI interactions, the identification of synthetic content, and certain deepfake disclosure obligations. It places obligations on providers and distributors of AI-generated human interaction to be transparent, disclose what is involved in an experience, provide machine-readable markings, and disclose when a deepfake is involved.

If your business uses general-purpose AI systems to generate synthetic audio, image, video, or text content, or to assess emotions or your biometric data, you will need to consider whether you should address your new obligation.

So, if it is not obvious to a reasonably well-informed, observant, and circumspect human, businesses should be clearer about how much AI is involved. The EU AI Act provides a framework and lexicon to do so.

Cyberthreats, deepfakes, and scams

The intention behind this law is for the good actors in industry to do the right thing under the EU AI Act. The policy intention is to help people recognise AI interactions and AI-generated content, thereby reducing risks of deception, manipulation, fraud, impersonation, and consumer deception. Article 50 is primarily a transparency provision, not a prohibition on fraud.

All of us, in all our dealings with digital and electronic systems, need to be alert to scammers who may still break the law through fraud, phishing, identity theft, or consumer protection violations, even if they disclose AI use. Article 50's purpose is to ensure people know when AI is involved and when content is synthetic.

The policy intention is to promote trust in business, and so to reduce the number of people who are suspicious of AI and how AI is being used.

Which change program do I need to include in my business plan?

In running your business, you need to think through your use of AI, your role in it, and the roles of your suppliers. In any case, businesses that take into account the EU AI Act should be able to identify if your firm is.

An AI provider: This is the company that builds or fine-tunes the underlying model and integrates it into a user-facing tool. It is responsible for making the machine-readable watermark or metadata directly into the generation pipeline.

An AI deployer: Businesses and end-users deploying AI tools must add clear, human-readable visible labels or disclosures (such as EU-compliant AI icons) when publishing realistic deepfakes or text concerning matters of public interest.

Whilst the headlines might be about watermarks, there is a deeper dive into knowing your business and the need to consider disclosure and transparency in the use of AI.

Britain is no longer in the EU; how will this affect us in the UK?

The EU is arguably leading the world in this governance; China has similar requirements; America has state rather than federal legislation, and the UK leaves most of this to common sense. It has long been recognised that serving your customer well, being open and honest with them, and delivering to their needs and aspirations is good business.

The EU's leadership in AI governance may have an effect similar to GDPR, creating a practical incentive for businesses operating internationally to adopt EU-aligned standards."

Applying good governance usually has the positive consequence of lower compensation and fines paid because bad practices are found out. The EU and its constituent countries remain the largest export market for UK businesses, and it is easier to apply a single governance system as the high watermark for all products and services rather than multiple standards.

Fines for failing to meet these transparency and watermarking mandates can reach up to €15 million or 3% of a company's total annual global turnover.

What if my business uses AI as an editor?

This has been thought about. The disclosure obligation does not apply where the AI-generated text has undergone human review or editorial control, and where a person holds editorial responsibility for the publication of the content.

So why should I deploy budget on this?

If you use AI, then you need to know how it impacts your business. Article 50 is not designed to eliminate fraud. It is designed to make digital interactions more trustworthy. Trust is a fundamental element of your brand. For UK businesses trading internationally, particularly those competing with firms already adopting EU standards, understanding where AI is used, being transparent about that use, and adopting consistent governance standards may prove just as important as technical compliance.