10 Responsibility and accountability

As GenAI becomes more integrated into decision-making, questions of accountability become critical.

When AI systems make mistakes – misdiagnosing a patient, generating biased decisions, or producing defamatory content – determining who is responsible is complex. However, this is exactly why it is so important to have clear lines of accountability – human accountability in the deployment of AI models (Floridi et al., 2018).

The legal position in the UK is still evolving to clarify liability. However, gaps remain. Ethical frameworks stress that developers, deployers, and end users must share responsibility.

Image generated using the AI prompt: Draw a responsibility flow chart for AI use in small organisations.

Show description|Hide description

A flowchart titled “RESPONSIBILITY FOR AI USE IN SMALL ORGANISATIONS” is displayed on a light beige background. At the top, a dark blue rounded rectangle labelled “AI USE” branches downward into three separate boxes, each representing a key role: “INDIVIDUAL EMPLOYEES” on the left, “IT” in the centre, and “MANAGEMENT” on the right. Each box is dark blue with white uppercase text, illustrating the distribution of responsibility for AI usage across these three groups within small organisations.

Image generated using the AI prompt: Draw a responsibility flow chart for AI use in small organisations.

Professionals, particularly in regulated sectors, cannot defer blame to algorithms. They remain accountable for ensuring outputs are fair, lawful, and appropriate.

Maintaining human oversight – the 'human-in-the-loop' – is essential to uphold autonomy, particularly when decisions have significant personal or legal consequences. Internal governance structures, documentation, and regular audits are critical tools for supporting accountability.