AI in Business: Who Is Liable When Technology Gets It Wrong?

Published Date: 07/23/2026
Source: Indwe Risk Services

---

It often starts with a simple prompt.

An employee asks an AI tool to draft an email. A consultant uses it to help prepare a client report. A finance team relies on it to analyse data. An HR department uses it to screen CVs or summarise applications.

Tasks that once took hours can now be completed in minutes, helping businesses to improve efficiency, reduce administrative workloads and support better decision-making.

But what if AI gets it wrong?

An AI-generated report may contain inaccurate information. A chatbot may provide incorrect information to a customer. Sensitive client information may be uploaded into an external AI platform. Or an employee may be persuaded by a convincing AI-generated voice message to authorise a fraudulent payment.

While the technology may have contributed to the mistake, responsibility may still rest with the business.

Although artificial intelligence can make businesses more efficient, it can also make mistakes faster. When an AI-generated recommendation leads to financial loss, confidential information is exposed or incorrect advice is shared with a client, the technology may have made the mistake, but the business may still be responsible.

This article explains where AI-related liability may arise, which insurance policies may be affected, and what practical governance measures can help businesses reduce emerging risks.

Key Context

Artificial intelligence is rapidly becoming part of everyday business operations. As adoption increases, businesses should understand both the opportunities AI presents and the risks that may accompany its use.

Key facts include:

• AI adoption is increasing across businesses of all sizes.
• Businesses commonly use AI for administration, customer service, finance, marketing, recruitment, reporting and decision support.
• AI systems can generate inaccurate information (“hallucinations”), biased outputs or incorrect recommendations.
• Employees may unintentionally upload confidential or personal information into AI platforms.
• AI can be exploited for phishing, deepfakes, invoice fraud and impersonation scams.
• AI-related losses may affect multiple insurance products, including cyber insurance, professional indemnity, directors’ and officers’ liability, crime and liability policies.
• South African businesses remain responsible for complying with the Protection of Personal Information Act (POPIA) when processing personal information.
• AI governance is becoming an important business risk management consideration globally.

Why AI Changes Business Risk

Artificial intelligence is a powerful business tool, but it is not an independent decision-maker. Like any other technology, the way it is used, monitored and applied remains the responsibility of the business.

This is a crucial distinction.

AI can analyse information, generate content and support decision-making in seconds. However, if inaccurate information is shared with a client, confidential data is exposed or an AI-assisted decision causes financial loss, the business may still be accountable for the outcome.

Human oversight therefore remains an essential part of responsible AI use. AI can support people, but it should not replace professional judgement, appropriate review processes or internal controls.

For organisations that provide advice or professional services, this is particularly important. If AI-generated information is included in a client report, recommendation or financial analysis without appropriate verification, the business may still face allegations of negligence, depending on the circumstances.

AI also introduces risks that extend beyond professional advice. Incorrect marketing content, biased recruitment recommendations, inaccurate customer communications, AI-generated financial reports, AI-generated emails containing incorrect advice, or the unauthorised disclosure of confidential information can all create operational, reputational and regulatory challenges. Employees using public AI tools without appropriate governance may also create unintended privacy, confidentiality or compliance risks. For directors and business leaders, AI governance is becoming an ever more important consideration. Understanding where AI is being used, what controls are in place and how risks are managed forms part of good governance as technology adoption continues to evolve.

Insurance should evolve alongside these operational changes. As businesses introduce new technologies and new ways of working, reviewing existing insurance arrangements can help identify whether emerging risks are appropriately considered as part of an organisation’s broader risk management strategy.

The role of an insurance adviser is no longer limited to traditional risks. As businesses adopt new technologies, insurance advice should also help identify how emerging operational risks may affect existing cover.

Practical Business Scenarios

Scenario 1: AI Generates Incorrect Professional Advice

A consultant uses AI to help prepare a report for a client. The report contains inaccurate information, which is not identified before it is shared. The client relies on the advice and subsequently suffers financial loss.

Practical Implications

Although AI assisted in producing the report, the business may still face allegations of negligence, depending on the circumstances.

Potential Insurance Considerations

Professional indemnity and errors and omissions cover may become relevant, subject to the policy wording, terms and conditions.

Scenario 2: Confidential Information Is Uploaded into an AI Platform

An employee uploads sensitive client documents into an external AI platform to summarise or analyse the information.

Practical Implications

Depending on the circumstances, this may create privacy, confidentiality and POPIA considerations, particularly where personal information is involved.

Potential Insurance Considerations

Cyber insurance and privacy liability may become relevant, depending on the nature of the incident and the policy wording.

Scenario 3: AI-Assisted Fraud

A finance employee receives a convincing AI-generated voice message instructing them to make an urgent payment. Believing the request to be genuine, the funds are transferred to fraudsters.

Practical Implications

As AI-generated scams become increasingly sophisticated, strong verification procedures and internal controls remain essential.

Potential Insurance Considerations

Crime insurance, cyber insurance and internal fraud prevention measures may all become relevant, depending on the circumstances.

Scenario 4: AI-Generated Marketing Creates Legal Issues

A business publishes AI-generated website or marketing content without reviewing it. The content contains inaccurate claims or copyrighted material.

Practical Implications

Publishing inaccurate or inappropriate content may create legal, contractual or reputational challenges for the business.

Potential Insurance Considerations

Media liability, legal costs and reputational considerations may become relevant, depending on the circumstances and policy wording.

Real-World Business Impact

AI-related incidents do not only create technology problems. They can have wider business consequences that affect operations, finances and customer relationships.

Depending on the nature of the incident, organisations may experience:

• Financial loss.
• Regulatory investigations.
• Data breaches.
• Loss of customer trust.
• Contract disputes.
• Reputational damage.
• Increased litigation risk.

As businesses become more reliant on AI, it is also important to recognise that insurance policies may not automatically respond to every new risk if changing business activities or exposures have not been appropriately disclosed.

Understanding how technology is changing an organisation’s risk profile before an incident occurs can help businesses make more informed decisions about governance, operational controls and insurance.

Artificial intelligence is creating new opportunities for businesses to improve efficiency, support decision-making and streamline everyday operations. At the same time, it is changing the way organisations are exposed to risk.

While AI may assist with generating content, analysing information or supporting business decisions, responsibility for how those outputs are used may still rest with the business. Appropriate governance, human oversight and well-considered risk management remain essential.

As AI becomes more deeply embedded in business operations, reviewing insurance should form part of a broader conversation about managing emerging risks.

Key Takeaways

• Artificial intelligence creates new business opportunities, but it may also introduce new liability exposures.
• Businesses remain responsible for how AI is used within their operations.
• Human oversight remains an essential part of responsible AI use.
• Multiple insurance policies may become relevant following an AI-related incident, depending on the circumstances and policy wording.
• Reviewing insurance before adopting new technologies is an important part of good governance and effective risk management.

Artificial intelligence is changing the way businesses work. It is also changing the way businesses are exposed to risk.

Understanding those changes before a claim occurs can help businesses make more informed insurance decisions.

The right insurance starts with a conversation.

It starts with Hello Indwe: https://indwe.co.za/contact.php

Indwe Risk Services is an authorised Financial Services Provider FSP 3425

Disclaimer: The information in this blog is for general information purposes only and does not constitute advice.