Businesses now rely on AI for daily operations, but this shift brings new legal and financial risks. Stanford University’s 2026 AI Index showed that 88% of surveyed organisations used AI, while documented incidents rose to 362 in 2025. Companies must manage these liabilities beyond just focusing on basic cybersecurity measures.
As businesses increasingly rely on AI for customer service, hiring, financial decisions and other operations, errors can create legal and financial consequences. Companies need to look beyond cybersecurity and assess their exposure to AI-related liability, governance failures and insurance gaps.
There are approximately 8.3 billion people in the world and an estimated 300-360 million businesses. Now imagine a world where every one of those businesses begins delegating more of its decisions to AI. An AI assistant answers a customer's complaint. Another reviews a contract. One recommends whether to approve a transaction, while another helps shape a product, assess a risk or make a hiring decision. The possibilities are enormous, and businesses are already moving from experimenting with AI to embedding it into everyday operations.
But there is a question that deserves as much attention as what AI can do: when an AI system gets something wrong, who is accountable for the consequences? For years, businesses have understood technology risk largely through the lens of cybersecurity: protect the systems, secure the data and prevent unauthorised access. Those safeguards remain essential. But AI introduces another dimension of risk. A system does not have to be hacked to cause harm. It can produce a misleading answer, make a flawed recommendation or take an action that creates financial loss, breaches a contractual obligation or damages customer trust.
The challenge, therefore, is not simply to secure AI systems. It is to understand and manage the business liability that can arise when they are used.
AI risk goes beyond cybersecurity
While cybersecurity remains a critical part of managing technology risks, it does not account for the full spectrum of risks that AI introduces. The scale of adoption makes this more pressing. Stanford University's 2026 AI Index reported that 88% of surveyed organisations used AI, while documented AI incidents rose from 233 in 2024 to 362 in 2025. The challenge is no longer simply getting businesses to adopt AI, but ensuring that risk management evolves alongside it.
This shift is visible in conversations with Indian startups. Over the last six months, we have observed a 45% increase in companies approaching us with questions about AI-related risks. Across more than 100 detailed conversations with founders and operators, four out of five startups seeking to understand their AI liability did not fully understand the extent of their exposure.
The problem is often a narrow understanding of insurance. A company may purchase substantial cyber cover while overlooking professional indemnity, directors and officers liability or general liability. Yet an AI-generated error that reaches a customer is fundamentally different from a data breach, and may require a different form of protection.
Understand where the liability sits
Consider a business deploying an AI chatbot to handle customer queries. If the chatbot provides incorrect information that a customer relies on, the company may still be held responsible. In 2024, Air Canada was held liable after its chatbot gave a customer incorrect information about a bereavement fare policy. The airline could not simply separate itself from the system's response.
The same principle applies internally. AI may draft contracts, screen job applicants, generate software code or make recommendations used in financial and operational decisions. Errors can expose businesses to claims, while decisions approved by leadership may also raise questions about governance and directors' responsibilities.
There is another layer to consider: AI that businesses never deliberately chose. Third-party software used for recruitment, customer relationship management or analytics may incorporate AI models. If these tools make errors or expose sensitive information, the business using them may still face contractual or legal consequences.
For startups, this exposure can extend across borders. Many Indian companies serve customers in the United States and Europe, where litigation and regulatory requirements can increase the potential cost of an incident. Our assessments have identified early-stage businesses with potential exposures exceeding $1 million, or approximately ₹10 crore, depending on their AI usage, sensitive data access and customer base.
Make risk management part of AI adoption
As AI becomes more deeply embedded in business operations, the question is no longer just how to deploy it securely, but how to remain accountable for its outcomes. Businesses cannot delegate responsibility along with decision-making. When an AI system makes a mistake, the consequences are ultimately borne by the business, its customers and, in some cases, its leadership.
This calls for a broader approach to AI risk -- one that brings together cybersecurity, governance, liability and appropriate insurance. The businesses best prepared for the next phase of AI adoption will not necessarily be those deploying it the fastest, but those that understand where their exposure lies and have safeguards in place when things go wrong.
Ultimately, responsible AI adoption is not just about what the technology can do, but whether businesses are prepared to take responsibility for what it does.
