
On August 2, 2026, the transparency obligations of the EU Artificial Intelligence Act (EU AI Act) take effect. Unlike other parts of the law that were postponed, this part was not — and it is relevant to far more organizations than it seems.
What exactly is required
The obligation has two parts. A provider — any organization that operates an AI system under its own name and brand, including an in-house chatbot — must clearly inform the user that they are talking to an artificial intelligence system and not a human. In addition, AI-generated content must be marked in a machine-readable format.
An enterprise user (deployer) — that is, an organization that uses an AI tool and did not necessarily develop it — must disclose deepfake content, and inform people when emotion recognition or biometric categorization technologies are applied to them.
The content marking obligation has a grace period until December 2, 2026, but only for systems placed on the market before that date. The disclosure obligation itself applies immediately.
Why this concerns an Israeli company
The EU AI Act applies based on where the user is, not where the company is. An Israeli company that operates a chatbot on its website and has users or customers in the European Union falls within its scope. This is also true for an internal support tool used by employees in Europe.
What was postponed — and what wasn't
Under the Digital Omnibus approved in May 2026, the heavy obligations for high-risk systems were postponed: standalone systems under Annex III (recruitment, credit scoring, education, law enforcement) will apply from December 2, 2027, and systems embedded in regulated products under Annex I — from August 2, 2028. The postponement is meant to give the European standardization committee time to publish standards.
The outright prohibitions — for example, emotion recognition in the workplace — have applied since February 2025. The AI literacy obligation, which requires organizations to train the employees who operate AI systems, is also already in force.
August 2 is also the date by which member states are supposed to appoint market surveillance authorities and receive enforcement powers. In practice most countries are not meeting it — so enforcement will begin gradually.
What This Means for Your Organization
- Check whether your website or systems include a chatbot that carries the company's name — including one deployed through an external vendor.
- If you have users in the European Union, make sure the statement "you are talking to an AI system" appears clearly at the start of the interaction, not in the terms of use.
- Contact your model provider and find out what it provides for machine-readable content marking. Almost no organization can implement this on its own.
- The AI literacy obligation is already in force — basic training for employees who use AI tools is a requirement, not a recommendation.
