Control is about having appropriate guardrails for AI use. It covers governance structures, policies, risk management, ethical guidelines, and accountability frameworks. Moving fast without guardrails creates liability. Good governance enables innovation by making it safe to experiment.
This dimension covers:
- • AI governance model and decision bodies
- • Policies and approval workflows
- • Risk management and compliance
- • Ethics principles and review processes
- • Accountability and human-in-the-loop
Cost of getting it wrong:
- • Regulatory violations and fines
- • Reputational damage from AI mistakes
- • Bias and fairness issues
- • Loss of stakeholder trust
- • Backlash that slows future AI adoption
Maturity Levels
Find your current level, then see what it takes to progress.
5 Proactive Governance enables innovation
"Our governance enables innovation rather than blocking it. Ethics is embedded in design processes, with continuous improvement."
You recognize this when:
- Governance that accelerates: Teams seek governance input because it helps, not because they have to
- Ethics by design: Ethical considerations are part of the initial design phase, not bolted on later
- Continuous improvement: Policies and practices evolve based on experience and changing capabilities
- Industry leadership: Others look to your organization for governance models and best practices
- Exemplary practices shared: You publish case studies and frameworks that benefit the broader community
↑ Sustaining excellence
- Share governance models externally
- Evolve practices as AI capabilities change
- Maintain balance between enabling and protecting
- Continue developing ethics expertise
4 Managed Risk-based governance
"We have risk-based governance with clear accountability. Human-in-the-loop requirements are defined, and ethics review is standard practice."
You recognize this when:
- Proportional governance: Light-touch for low-risk uses, thorough review for high-stakes decisions
- Clear accountability: Named individuals responsible for AI decisions in their domains
- Human-in-the-loop defined: Each use case has explicit rules for when humans must review or approve
- Standard ethics review: New AI projects automatically go through ethics assessment
- Continuous monitoring: Systems in place to catch issues early, before they become problems
↑ Moving to Level 5
- Make governance an enabler (not blocker)
- Embed ethics in design process
- Develop continuous improvement practices
- Build industry thought leadership
3 Defined Governance body established
"We have a governance body with documented policies and approval workflows. Ethics principles are formally stated."
You recognize this when:
- Governance body meets: A committee or council reviews AI initiatives regularly
- Documented policies: Written guidelines are accessible to everyone who needs them
- Approval workflows: New AI initiatives go through a defined process before deployment
- Ethics principles stated: The organization has articulated what it believes about responsible AI
- Clear decision processes: People know how AI decisions get made and who makes them
↑ Moving to Level 4
- Implement risk-based governance
- Define clear accountability framework
- Establish human-in-the-loop requirements
- Make ethics review standard for new projects
2 Reactive Policies created after issues
"We have basic AI policies, created in response to incidents. There is minimal awareness of ethical considerations."
You recognize this when:
- Policies from incidents: Rules exist because something went wrong, not from proactive planning
- Compliance-driven: Legal or compliance team created rules after an issue was raised
- Limited awareness: Most people don't know AI policies exist or where to find them
- Ethics in passing: Ethical considerations mentioned occasionally but not systematically addressed
- Reactive stance: Waiting for problems to emerge rather than preventing them
↑ Moving to Level 3
- Establish a governance body
- Document comprehensive policies
- Create approval workflows
- State ethics principles explicitly
1 Absent No governance exists
"No AI governance exists. Decisions about AI use are made ad hoc, with no one thinking about ethics or policy."
You recognize this when:
- No ownership: No one owns AI governance—it's nobody's job
- Case-by-case decisions: Each AI initiative handled individually with no framework
- Ethics absent: Nobody has raised ethical questions about AI use
- Accountability unclear: When something goes wrong, there's confusion about who's responsible
- Invisible risk: Risk accumulates without anyone tracking or managing it
↑ Moving to Level 2
- Create basic AI usage policies
- Assign someone accountable for AI decisions
- Start ethics conversations
- Document initial guidelines
Common Patterns
Governance as blocker — Policies so restrictive that innovation stalls. Need to shift from gatekeeping to enabling.
Ethics afterthought — Building first, asking ethical questions later. Need to embed ethics in design process.
Accountability gap — Nobody clearly responsible when things go wrong. Need named individuals for AI decisions.