What Just Happened on August 2?
The European Union has officially started enforcing the transparency obligations under Article 50 of its comprehensive AI Act. This isn't the entire law—many provisions, especially for high-risk systems, are phased in later—but Article 50 is the part
that affects almost every company creating or using AI to produce content. The rules apply to any business whose AI outputs are used in the EU, regardless of where the company is based. This means providers of AI systems that generate text, images, audio, or video must now ensure their outputs are marked in a way that is detectable as being artificially generated. For AI that interacts directly with people, like chatbots, disclosure is also required unless it's already obvious.
Decoding 'Machine-Readable Markers'
At its core, a machine-readable marker is a digital signal embedded into content that allows software to identify it as AI-generated. Think of it as a sophisticated, invisible watermark. This isn't just about a visible label that says "Made with AI." The EU's rules require a technical solution that is robust, reliable, and interoperable. The leading industry standard emerging to meet this demand is C2PA (Coalition for Content Provenance and Authenticity), which creates a kind of digital birth certificate for content, logging its origin and any changes made. While the AI Act doesn't explicitly name C2PA, the European Commission's guidelines point to it as a prime example of a compliant technology. The goal is to create a reliable way for platforms and users to trace content back to its source and verify its authenticity.
Why This Is a 'Practical' Requirement
The headline's use of "practical requirement" is key. While Article 50 lays down a legal mandate for AI providers, the ecosystem dynamics make adoption a business necessity for everyone else. For providers of generative AI, the obligation is direct: mark your outputs or face fines of up to €15 million or 3% of global turnover. For companies that use these tools (deployers), the pressure is indirect but just as strong. Using unmarked AI content for public-interest matters or creating deepfakes without disclosure also falls foul of the rules. Furthermore, to mitigate legal risk and maintain user trust, platforms are increasingly likely to demand marked content. The EU has also established a voluntary Code of Practice, and signatories who follow it are given a degree of presumed compliance, creating a powerful incentive to adopt best practices like C2PA.
The Global Ripple Effect: The 'Brussels Effect'
This isn't just a European affair. The AI Act is a textbook example of the "Brussels effect," where EU regulations become global standards because companies find it easier to apply one high standard across all their operations rather than create different versions for different markets. Major AI developers will build these marking capabilities into their foundational models. To stay competitive and compliant, companies worldwide, including those in India, will need to integrate these new workflows. California's own AI transparency law notably came into effect on the very same day, August 2, mirroring the EU's focus on provenance and machine-readable markers, signaling a growing international consensus.
What This Means for Indian Tech
For India's massive IT and business process outsourcing (BPO) sectors, the EU AI Act is a direct and immediate concern. NASSCOM has previously estimated that over 1,200 Indian IT firms have EU clients, and any AI systems or content they provide for that market now fall under these rules. An Indian HR-tech company providing an AI-powered CV-screening tool to a client in Germany, for instance, is now a 'provider' under the Act. Similarly, a digital marketing firm in Bengaluru using generative AI to create ad campaigns for a French brand is a 'deployer'. This necessitates a rapid upskilling in technical and regulatory compliance, and a shift towards creating hybrid roles that blend legal expertise with AI development. Indian companies must now audit their AI usage, invest in compliant technologies, and prepare for a new era where content provenance is not just good practice, but a prerequisite for market access.
