In this webinar we tackle one of the most pressing — and often avoided — questions in the enterprise AI conversation: when AI makes mistakes, who is responsible? This is not merely a philosophical question. It has direct legal, reputational, and operational consequences for every organisation deploying AI systems.
The Accountability Gap in AI
The rapid deployment of AI systems has outpaced the development of clear accountability frameworks. When an AI chatbot gives a customer incorrect medical advice, when an AI hiring system discriminates against a protected group, or when an autonomous vehicle causes an accident — the question of liability is rarely straightforward.
Ethical AI governance requires companies to think carefully about:
- The chain of responsibility from AI developer to deployer to end user
- How to maintain meaningful human oversight without eliminating the efficiency benefits of automation
- What documentation and audit trails are required to demonstrate responsible deployment
- How to respond when things go wrong — including incident response and remediation
Three Real-World Scenarios
The webinar examines three concrete scenarios where AI accountability becomes critical:
Scenario 1: The Customer Service Chatbot That Gives Wrong Information
A chatbot incorrectly tells a customer their product is covered under warranty when it is not. The customer relies on this information, and a dispute follows. Who is liable — the AI provider, the company deploying the chatbot, or the customer for trusting an automated system?
Scenario 2: The AI-Assisted Recruitment Tool
An HR AI system trained on historical promotion data systematically ranks male candidates higher than female candidates. Candidates are rejected based partly on AI recommendations. Who is responsible for the discriminatory outcome?
Scenario 3: The AI Agent That Reveals Sensitive Information
A customer service AI, through a prompt injection attack, reveals information about another customer. What obligations does the company have? Who bears liability?
The EU AI Act and Liability Framework
The EU AI Act and the proposed EU AI Liability Directive provide the emerging regulatory framework for AI accountability in Europe. Key provisions relevant to enterprise AI deployments include:
- High-risk AI systems must maintain detailed technical documentation
- Deployers of high-risk AI must conduct fundamental rights impact assessments
- Users harmed by AI systems may be entitled to compensation — the burden of proof is shifting towards AI deployers
- Companies must be able to demonstrate meaningful human oversight was in place
Building an Ethical AI Governance Framework
The webinar presents a practical ethical AI governance framework that companies of any size can implement:
- AI inventory: document every AI system in use and its risk classification
- Impact assessment: evaluate potential harms before deployment
- Human oversight protocols: define when and how humans review AI decisions
- Incident response plan: establish procedures for when AI causes harm
- Audit and monitoring: continuous review of AI system behaviour and outcomes
FAQ: Ethical AI and Accountability
Is a company legally liable for mistakes made by its AI chatbot? Under EU law, companies deploying AI systems are generally considered responsible for their outputs. The AI Act and AI Liability Directive are strengthening this liability framework significantly.
How can companies reduce AI liability risk? Key steps include: using RAG to ground AI in verified information, maintaining human oversight for high-stakes decisions, keeping detailed audit logs, and having clear user disclosures about AI interaction.
What should companies do immediately after an AI-caused incident? Immediately suspend the AI system if the risk is ongoing, document what happened, notify affected parties as required by GDPR and sector-specific regulations, and conduct a root cause analysis before redeployment.