Europe's AI Transparency Rules Go Live, and India Is Watching
Brussels has switched on the world's first hard rules forcing firms to label chatbots and deepfakes. For Indian software exporters, the clock has already started.
The News
The European Union's landmark transparency obligations under the AI Act came into force on 2 August 2026, requiring companies to tell people plainly when they are talking to a machine and to flag content that a model has generated or altered. The rules, reported by The Verge, target chatbots, synthetic media and deepfakes across every service that reaches European users.
To smooth the transition, Brussels is not leaving firms to invent their own warning systems. The European Commission has published a voluntary Code of Practice on the marking and labelling of AI-created content, complete with ready-made icons that creators and publishers can drop into their products rather than designing bespoke disclosures. Accompanying guidelines clarifying the scope and exceptions were issued on 20 July 2026.
In practice, a deepfake video or an AI-written article on a matter of public interest must now carry an identifiable marker. An interactive assistant must announce itself. The obligations sit inside the same regulation whose most serious breaches can draw penalties running into tens of millions of euros or a slice of global turnover, giving the labelling requirements real teeth.
Why It Matters
This is the first time a major jurisdiction has moved AI disclosure from guidance to enforceable law, and it establishes a template others will copy. The pattern echoes the arrival of the General Data Protection Regulation in May 2018, when a single European statute quietly rewrote privacy practice for firms on every continent that wanted access to the bloc's consumers.
The reach is deliberately extraterritorial. A company does not need an office in Frankfurt or Paris to fall within scope; it only needs its output to land in front of an EU user. That turns a regional rulebook into a de facto global standard, the so-called Brussels effect, and it shifts the burden of proof onto builders to show their synthetic content is honestly marked.
The timing also matters. With generative tools now embedded in newsrooms, customer support and political messaging, regulators are betting that visible provenance can slow the spread of convincing fakes before trust erodes further.
Indian Angle
For India, this is not a distant European story. The country's technology services giants, from TCS to Infosys and Wipro, build and operate AI systems for European clients, and those deployments now inherit the labelling duty. Compliance engineering, audit trails and disclosure logic become billable work, but also a fresh source of contractual risk if a client's chatbot goes unmarked.
Home-grown model builders such as Sarvam AI and Ola's Krutrim face the same gate the moment they court European users or partners. Any Indian firm selling a generative product into the single market must now treat provenance marking as a shipping requirement, not an afterthought.
Domestic policy is drifting the same way. India's Ministry of Electronics and Information Technology has already floated proposed amendments to the IT Rules that would require platforms to label synthetically generated information, and the DPDP Act framework is tightening consent norms. Indian firms that align early with Europe's standard may find they have quietly future-proofed themselves against Delhi's own coming rules.
FAQ
When did the rules take effect?
The transparency obligations became enforceable on 2 August 2026. The Commission's clarifying guidelines were published shortly before, on 20 July 2026, and a voluntary labelling Code of Practice is available for firms to adopt.
Do the rules apply to Indian companies?
Yes, if their AI output reaches users in the European Union. The AI Act applies based on where the system is used, not where the company is based, so Indian exporters and startups serving EU customers are firmly in scope.
What content must be labelled?
Chatbots must disclose they are not human, and generative content must be identifiable. Deepfakes and AI-written text published to inform the public on matters of public interest face specific marking requirements under the new obligations.
Is India planning similar rules?
MeitY has proposed amendments to the IT Rules that would require labelling of synthetically generated content, signalling that a comparable Indian framework is taking shape alongside the DPDP Act.
This story was reported by The Verge. Read the full original coverage at The Verge.