Who is accountable for AI?

Okay … putting my new AI ethics and compliance hat on, let’s start with a simple question.

In your organisation, who is ultimately accountable for the use of AI? If, tomorrow, your business was found to have used AI in an unethical way, where would that responsibility sit?

For example, imagine your HR team have been using an AI tool to screen candidates. On the surface, that appears quite efficient and objective. But over time, it might have unintentionally filtered out strong candidates based on gender, race or another demographic factor. If that became public and impacted trust in your organisation, who would be held accountable?

Or consider a different scenario. A website update introduces AI generated content that hasn’t been properly reviewed and it turns out to be inaccurate or misleading. We’ve all heard about hallucinations and AI generated “slop”, so it’s not a far-fetched situation. Again, where does responsibility lie?

A third scenario might be your Finance team using an AI tool to support credit limit decisions or payment risk scoring. Over time, some customers have become disadvantaged due to flawed outputs that weren’t properly sense checked.

Is the accountability with:

• The technology function?

• Individual department heads?

• The executive team?

• The board?

• Or ultimately, the CEO?

I’m genuinely interested in how others see this.

I recently attended an event exploring the opportunities and challenges of scaling AI, with contributors from organisations such as GCHQ, NCSC, some well-known multinationals, alongside a fairly diverse audience. This exact question came up and, interestingly, there wasn’t a clear or consistent answer.

That, in itself, is quite telling.

Link to original LinkedIn post