Business Leaders Don't Have to Wait for Regulation
Governance is a decision you can make now.
Business leaders are waiting.
They’re waiting for the government to set the rules. Waiting for the technology companies to figure it out. Waiting for a clearer picture before they commit to anything.
The waiting is understandable. It’s also a mistake.
Who actually owns AI responsibility?
Someone on a business message board recently asked a question worth sitting with: who is ultimately responsible for AI?
The honest answer is all of them: government, technology companies, communities and business leaders. But not equally, and not in the same way.
Government sets the outer boundaries. Technology companies design the systems. Communities shape expectations. Company leaders make the operational decisions that determine how AI actually gets used inside an organization:
• Which data goes in.
• Which uses get approved.
• Which outputs get reviewed before they affect a real decision.
• Who is accountable when something goes wrong.
That’s the layer of accountability that matters most right now, and it’s the one most organizations are leaving unstructured.
The South Africa lesson
In April 2026, South Africa’s government withdrew its draft national AI policy after it was found to contain fake citations: references to sources that don’t exist, generated by AI.
It isn’t the only case. Fabricated citations have already turned up in court filings and professional reports.
The problem wasn’t that AI made things up. That’s known, documented behavior. The problem was that the output made it into an official government document without enough checking. No one with authority caught it in time.
For businesses, the parallel is uncomfortable. How many AI-generated outputs are moving through your workflows right now with no formal review? How many decisions are being shaped by AI-produced analysis that no one has independently checked?
The risk isn’t that AI is dangerous in the abstract. It’s that AI is being used quietly, without clear ownership, without a structure for accountability, and without anyone formally responsible for the result.
AI moves accountability. It doesn’t remove it.
This is the point most business leaders are missing.
When your organization uses AI to make or influence a serious decision, the accountability doesn’t transfer to the technology. It stays with you. AI changes where the question sits, not whether the question exists.
• Who reviewed the output?
• Who approved the use?
• Who checked the source material?
• Who decided the risk was acceptable?
• Who is accountable when the system is wrong?
If you can’t answer these questions for the ways you use AI today, you don’t have a technology problem. You have a governance gap.
Governance isn’t a regulation you wait for. It’s a decision you make.
Where to start
You don’t need to build a complete AI governance program this quarter. You need to answer three questions:
1. What AI tools are in use inside your organization, formally or informally?
2. Which decisions or outputs are those tools influencing?
3. Who is accountable for reviewing those outputs before they affect a real outcome?
If you don’t have clear answers, that’s your starting point. Not a technology audit. A governance conversation.
AI will keep improving, and its capabilities will keep expanding. The accountability question won’t go away. It will get harder as the tools become more embedded in how decisions get made.
Organizations that build governance now will be able to use AI with confidence, not just caution. That’s the difference between organizations that adopt AI well and organizations that end up in the headlines for the wrong reasons.
Fellowship Intelligence works with leadership teams on AI strategy, governance and leadership. If this raised questions about how your organization governs AI, contact us.
Sources
Engineering News, “South Africa withdraws AI policy due to fake AI-generated sources,” April 28, 2026: https://www.engineeringnews.co.za/article/south-africa-withdraws-ai-policy-due-to-fake-ai-generated-sources-2026-04-28
