Nano Tools for Leaders—a collaboration between Wharton Executive Education and Wharton’s Center for Leadership and Change Management—are fast, effective tools that you can learn and start using in fewer than 15 minutes, with the potential to significantly improve your success and the engagement and productivity of the people you lead.

Goal

Deploy AI systems with confidence by ensuring they are fair, transparent, and accountable, which minimizes risk and maximizes long-term value.

Nano tool

As organizations accelerate their use of AI, the pressure is on leaders to ensure that these systems are not only effective but also responsible. A misstep can result in regulatory penalties, reputational damage, and loss of trust. Accountability must be designed in from the start, not bolted on after deployment.

Action steps

1. Define clear use cases and boundaries

Specify a well-understood purpose for each AI system. Document what the AI should and shouldn’t do, including red lines (e.g., no use of facial recognition in sensitive contexts). Link the use case directly to business goals and ethical commitments.

2. Establish a governance framework

Form a cross-functional governance board or policy that includes leaders from legal, risk, ethics, and operations—not just data science. Set standards, review high-impact use cases, and update guardrails regularly as technologies and risks evolve.

3. Assign human accountability

Designate a person or team responsible for the AI system’s behavior and effect—beyond technical oversight. Ensure that this group has legal, ethical, and operational authority, as well as clear pathways for raising and addressing concerns in real time.

4. Ensure explainability

Use AI models that can be explained to nonexperts. Communicate what the model does, what it’s trained on, and why it made specific decisions. If a decision can’t be explained, it can’t be trusted.

5. Test for bias and harm

Regularly audit AI outputs for unintended bias or discriminatory effect, aligned with organizational values and risk tolerance. Simulate edge cases using synthetic or real-world data, and embed fairness checks throughout the development life cycle.

6. Document and communicate decisions

Maintain clear records of how the AI was trained, tested, deployed, and updated. Share high-level information with stakeholders and employees to build trust, and continue to evaluate systems post-deployment.

How organizations use it

The following examples detail current responsible AI (RAI) governance and activity.

JP Morgan

Extensive, visible RAI activity throughout the firm; head of AI policy reports to the CEO; chief information security officer released a public letter to third-party suppliers (April 2025); dedicated RAI governance within model risk; 20-plus staff (not counting other RAI functions); in-house RAI development and research.

Salesforce

The Office of Ethical and Humane Use (now part of its broader RAI efforts) was established in 2018 to guide product development in line with ethical principles and to proactively tackle emerging ethical and safety challenges associated with technology—especially AI. Office includes ethicists, policy experts, researchers, and technologists who work across the company to assess risk and build trust; RAI is incorporated in enterprise goal-setting.

Mastercard

Established an AI Governance Council to oversee AI initiatives through cross-functional review, human oversight, and ethical guardrails; formalized data and tech responsibility principles, including privacy, transparency, accountability, fairness, and inclusion as core pillars; recently partnered with Quebec Artificial Intelligence Institute (Mila) to advance RAI research—particularly in bias testing and mitigation—and is bringing those findings into real-world AI deployments.

Published Sept. 23, 2025, by Knowledge at Wharton.