Recent media coverage of Deloitte’s government report, later found to contain fabricated quotes and references, has prompted serious questions about the responsible use of artificial intelligence.
But let’s be clear: this wasn’t a failure of artificial intelligence alone. It was a failure of governance. Humans made the decisions about how AI was used, how its output was verified, and what was delivered to a client. The technology didn’t go rogue, oversight did.
AI doesn’t remove accountability. It demands more of it. And when governance slips, even the smartest systems can expose an organisation to risk.
Governance Is the Foundation for Responsible AI
AI governance provides the structure and oversight needed to ensure that artificial intelligence is used ethically, transparently and safely. It defines who can use AI, for what purpose, and under what conditions. It also sets out expectations for accuracy, disclosure, and review.
Without these foundations, organisations risk exposing themselves to reputational, legal and operational harm. Just as poor recordkeeping can undermine compliance and trust, ungoverned AI use can produce outputs that are unreliable or misleading, sometimes without anyone realising until too late.
Lessons from the Deloitte Case
The Deloitte example offers valuable lessons for all organisations adopting or experimenting with AI technologies:
-
Understand how AI is being used
In this case, a draft version of the report reportedly contained AI-generated references and a fabricated legal citation. This underscores the importance of knowing where and how AI tools are being applied within your organisation, whether in content creation, data analysis or report drafting. AI should assist human work, not replace professional judgement. -
Human review must evolve
Traditional review processes are not enough when AI is involved. Organisations need robust verification steps to check facts, sources and references before information is shared externally. Establishing clear roles and checklists for AI-assisted outputs helps maintain accountability and quality.
-
Transparency builds trust
One of the key concerns raised was that AI use was not disclosed upfront. Transparency about when and how AI has been used, both internally and externally, supports informed decision-making and helps preserve stakeholder confidence. -
Policy and process are essential
AI tools are increasingly easy to access, but not all are suitable for sensitive or high-stakes work. An organisational AI policy should outline approved tools, acceptable use cases, prohibited applications, and responsibilities for review and oversight.
Embedding AI Governance in Practice
AI governance is not just about managing risk. It’s about enabling innovation safely. When clear policies, procedures and review processes are in place, AI can be used with confidence and consistency.
Strong governance helps organisations:
- Manage reputational and compliance risks
- Ensure accuracy and accountability in outputs
- Support ethical and transparent use of AI tools
- Strengthen decision-making and oversight
Building Confidence Through AI Governance
At Recordkeeping Innovation, we help organisations use AI safely and effectively through governance frameworks that promote accountability, transparency and explainability. Our work focuses on integrating AI governance into existing information management practices, ensuring that innovation is supported by clear oversight and ethical use.
We support clients through:
- AI governance frameworks and policy development
- Risk and readiness assessments
- Integration of AI governance with existing information and privacy frameworks
- Review of current processes to ensure responsible adoption
Final Thought
Artificial intelligence offers enormous potential, but without governance, it can just as easily create new risks.
The Deloitte incident serves as a reminder that technology should never outpace accountability. With clear frameworks, strong oversight and a culture of transparency, organisations can harness the benefits of AI while safeguarding accuracy, trust and credibility.
If your organisation is ready to strengthen its approach to AI governance, contact the Recordkeeping Innovation team today.