Europe's AI Rulebook Moves from Law to Action: A New Phase of AI Regulation Begins on 2 August 2026

On 2 August 2026, the global technology landscape enters a new era as the European Union officially begins enforcing the Artificial Intelligence (AI) Act. From 2 August 2026, the European Union begins enforcing key provisions of its Artificial Intelligence (AI) Act, marking an important step in how AI is regulated. The European Commission's AI Office and national authorities can now oversee, test, and regulate advanced AI models across the EU. The rules are designed to ensure AI systems are safe, transparent, and trustworthy, while setting a global benchmark for AI governance.

A Graduated, Risk-Based Approach

The new framework divides AI systems into different risk categories, with stricter rules for technologies that could have a greater impact on people's rights, health, and safety. Applications presenting an unacceptable risk face an outright ban. The ban covers AI used to predict crimes, gather facial images without consent, and monitor people's emotions in workplaces and schools.

The EU AI Act introduces a risk-based approach to AI regulation, with stricter rules for systems that have a greater impact on people.

AI systems that could have a greater impact on people's lives must meet stricter safety and transparency requirements before they can be used. Companies developing or using these AI systems must keep detailed technical records, use reliable data, and ensure that people can review and oversee important AI decisions. To help enforce these rules, the EU has established an AI Office in Brussels. The office is responsible for monitoring advanced AI systems and addressing serious risks, including illegal content and cyberattacks targeting public infrastructure.

The Transparency Mandate and Content Labelling

One of the main goals of the new rules is to make AI more transparent. Under the new regulations, customer service chatbots and interactive digital assistants must explicitly inform users that they are interacting with a machine rather than a human being.

The legislation also requires AI-generated images, videos, audio, and other content—including deepfakes—to be clearly labelled so that people know they were created or altered using AI. Developers must also include digital markers that help identify AI-generated images, videos and audio. To facilitate this complex technical transition, the European Commission recently published a voluntary Code of Practice, which has already been signed by over 180 organisations, aiming to operationalise and standardise these digital marking requirements.

Financial Penalties and Future Timelines

Companies that fail to comply with the AI Act could face significant financial penalties. Entities found engaging in prohibited practices face administrative fines of up to €35 million or 7 percent of their global annual turnover, whichever figure is higher. Other significant violations can incur penalties of up to €15 million or 3 percent of worldwide revenue. The EU has also launched a secure online platform where people can confidentially report companies they believe are violating the AI Act.

The EU AI Act introduces strict penalties, phased implementation, and new compliance requirements for businesses, with implications that extend beyond Europe.

Moving forward, the legislation will continue its phased implementation. A strict ban on systems generating sexualised deepfakes will take effect in December 2026, followed by the application of comprehensive rules for stand-alone high-risk systems in December 2027, and finally, regulations for high-risk tools embedded inside other products in August 2028.

Implications for the Indian Technology Sector

Although the AI Act is an EU law, its impact extends well beyond Europe. Indian technology companies that develop AI systems or provide AI-enabled services to customers in the EU will also need to comply with these rules. The legislation applies to any provider or deployer whose automated outputs are used within the European market, regardless of whether the company maintains a physical office in the bloc.

Industry estimates from NASSCOM indicate that over 1,200 Indian technology companies currently serve European clients potentially affected by the regulation. Indian enterprises developing tools such as automated resume screening software, predictive credit scoring systems, or machine learning analytics for European businesses must now adhere to the strict technical documentation and conformity assessment standards demanded by the new laws. Consequently, achieving compliance with this European framework is no longer optional; it has become a mandatory operational requirement for Indian firms aiming to sustain and expand their footprint in the lucrative digital economy.