Small and medium-sized businesses across the European Union that use AI on websites, in customer support, or for marketing content face new transparency duties from August 2, 2026, after Article 50 of the EU AI Act began applying.
The rules, clarified in final European Commission guidance published in late July, require providers and deployers of certain AI systems to tell people when they are interacting with AI and to disclose or mark some AI-generated or manipulated content. For SMEs, the most immediate impact is likely to be on customer-facing websites, lead-generation chat widgets, support bots, synthetic visuals, AI-made audio or video, and public-facing informational content.
The change is already live. The Commission says Article 50 transparency obligations apply from August 2, with only a limited grace period for certain marking and detection obligations on generative AI systems that were already on the market before that date.
For business owners, the practical message is straightforward: AI can still speed up marketing, support and sales, but hidden AI on a website is now a compliance and trust risk.
Chatbots must identify themselves from the start
One of the clearest obligations concerns AI systems that interact directly with people. According to the Commission FAQ, users should be informed when they are interacting with an AI system “from the start of the first interaction” and in a clear and distinguishable way.
That matters for companies using website chatbots, customer-service assistants, sales qualification widgets or lead-generation pop-ups powered by AI. TechRadar has framed the issue directly around SMEs, noting that businesses using off-the-shelf AI widgets for customer service or AI tools for marketing content are “deploying AI.”
For a small business, this means a chatbot should not wait until midway through a conversation to reveal that it is automated. The disclosure should appear before the customer begins using it, or at the very start of the exchange, in language that is easy to notice and understand.
A practical website message could be simple: “You are chatting with an AI assistant.” The key is that the information must be clear and distinguishable.
AI images, video, audio and text may need labels or marks
Article 50 also affects synthetic media and some AI-generated text. The Commission’s quick facts page says providers must apply machine-readable marks to synthetic AI content. Deployers must clearly disclose deepfakes and certain AI-generated text on public-interest topics, unless there is human review or editorial control.
For SMEs, this can reach beyond obvious “deepfake” scenarios. Marketing teams increasingly use AI to create images, video, audio, voices, avatars, social media assets, website illustrations, ad creative and written content. Where realistic AI-generated or manipulated content could be taken as real, disclosure duties may be triggered.
The rules are especially relevant to public-facing informational content. Businesses using AI to draft articles, explainers, reports, email campaigns or other materials on public-interest topics should pay attention to the human review and editorial control point in the Commission guidance. Human oversight is not only a compliance safeguard; it can also reduce reputational risk when customers are relying on the information.
Penalties are significant, even with SME proportionality
The Commission’s quick facts page lists penalties of up to €15 million or 3% of total worldwide annual turnover for breaches of the transparency rules, with proportionality for SMEs and small mid-caps.
That proportionality is important, but it does not make the rules optional. Many smaller firms adopted AI quickly because it is easy to add a chatbot, generate product visuals, automate marketing emails or draft web copy. The same low barrier to adoption can create a compliance gap if no one in the business has mapped where AI is being used.
More EU businesses are using AI
The compliance audience is expanding fast. Eurostat says 20% of EU enterprises with at least 10 employees used AI in 2025, up from 13.5% in 2024. The most common uses included analysing written language, generating images, video or audio, and generating written or spoken language.
Those uses closely match the areas now affected by transparency duties: language tools, synthetic media and customer-facing AI interactions. For SMEs, the rules are not only a legal issue for technology departments. They are also a marketing, sales, customer support and content operations issue.
Lithuania has a strong local stake
Lithuanian businesses are part of this shift. The European Commission’s 2026 Digital Decade country report says AI adoption in Lithuania’s business sector is now slightly above the EU average. Lithuania’s Economy and Innovation Ministry says €48 million has been earmarked for SMEs developing AI solutions, with additional investment planned.
That combination — rising adoption and new transparency duties — makes the Article 50 deadline particularly relevant for Lithuanian SMEs. Companies investing in AI-powered services, content workflows and customer-facing tools now need to make sure transparency is built into those systems from the start.
What SMEs should fix now
Businesses do not need to stop using AI, but they should review how it appears to customers and the public. A practical first step is to audit every AI tool used on the website and across marketing, support, sales, SEO, advertising, email and content production.
The most urgent fixes include:
- Audit all AI tools. List website chatbots, support assistants, lead-generation widgets, AI writing tools, image generators, video tools, voice tools, avatars, marketing automation and content production systems.
- Add chatbot disclosure. Make clear before or at the start of the first interaction that the user is dealing with an AI system.
- Label realistic synthetic media where needed. Review AI-generated or manipulated images, video, audio, voices, avatars and deepfake-style promotional content that users could mistake for real.
- Review AI-generated text workflows. For public-facing informational content, especially on public-interest topics, use human review and editorial control.
- Ask vendors for written confirmation. SMEs using third-party tools should ask how vendors handle AI marking, watermarking and Article 50-related compliance.
- Use transparency as a trust message. Clear disclosure can reassure customers that AI is being used responsibly rather than hidden from them.
The immediate risk for SMEs is not that AI use is banned. The risk is that AI is embedded into customer-facing systems without disclosure, review or vendor checks.
Article 50 turns transparency into an operational requirement. For businesses that rely on AI to move faster, the next task is to make that AI visible, labelled where required, and supported by human oversight where it matters.
