Artificial Intelligence | EU Drafts Rules: Who Pays When AI Causes Harm?

Friday, September 11, 2026.

Brussels just made a big move on how we deal with Artificial Intelligence. The European Commission, the main office of the European Union, today released its draft plan for something called the AI Liability Directive. Simply put, these are rules to decide who is legally responsible when AI systems cause harm.

Think about it: if an AI-powered self-driving car gets into an accident, whose fault is it? The car maker? The software developer? The company that maintains it? These new rules aim to answer such questions. They want to make it easier for people and businesses to get help if they suffer damages because of AI. This could mean getting money back or holding the right party accountable.

This directive follows months of discussions. These talks focused on how the broader AI Act — a separate, bigger set of AI rules also being worked on by the EU — would affect things. While the AI Act focuses on making sure AI systems are safe before they are used, this new directive steps in when things go wrong.

Global Race for Rules

The EU isn't the only one trying to figure this out. Just recently, the U.S. Department of Commerce suggested new safety testing standards for high-risk generative AI models. Generative AI is the kind that creates new things, like text or images, similar to what you see in ChatGPT.

Also, digital ministers from the G7 nations — a group of the world's biggest economies — met in Tokyo. They agreed that countries need to work together more on AI rules. This is important to avoid a confusing mess where every country has different, clashing rules. They especially talked about keeping our data private and making sure AI decisions are transparent, meaning we can understand how AI reaches its conclusions.

What This Means for India

Here in India, our government's Ministry of Electronics and Information Technology (MeitY) has been discussing AI regulation too. While the EU's rules won't apply directly to us, they set an important example. Many Indian startups and big tech companies either sell their AI products in Europe or use European-developed AI tools.

If an Indian company's AI system causes harm to a user in Europe, these new EU rules might apply. This means Indian developers need to pay attention to global standards. It's a reminder that India, too, will need its own clear rules. Our rules must protect users but also allow our AI industry to grow.

The catch, of course, is that writing laws for AI is tricky. The technology changes fast. What counts as 'damage' from AI isn't always clear. For now, the details of how 'easy recourse' will truly be under these new EU rules are still being worked out.