Artificial intelligence can help public institutions process information, deliver services, and make decisions. But as AI systems become more capable, government leaders face a question technology cannot answer: Who remains accountable when those systems influence decisions affecting the public?
For Tasha Austin-Williams, a lecturer at Cornell University’s Jeb E. Brooks School of Public Policy, the concern goes beyond whether AI produces inaccurate results.
Why Human Judgment Matters in Public-Sector AI
As AI becomes part of an organization’s operations, employees need more than technical familiarity. They must understand when to question a recommendation, investigate its limitations, or make a decision independently.
“You don’t need to be a technologist to lead in the AI era, but you cannot delegate your leadership responsibilities to an algorithm.”
For Dr. Austin-Williams, effective oversight means preparing employees to exercise judgment rather than simply encouraging them to adopt new tools.
That distinction matters when AI informs decisions about public services, resource allocation, and the communities government institutions serve.
Assessing Organizational Readiness for AI
Before adopting AI, public-sector leaders must evaluate whether their organizations are prepared to use it responsibly.
Dr. Austin-Williams identifies data quality, governance, workforce capabilities, organizational culture, and alignment with institutional priorities as essential considerations.
“AI doesn’t automatically correct institutional weaknesses; it can amplify them.”
A sophisticated system cannot compensate for unclear accountability, unreliable data, or employees who have limited support to evaluate its recommendations.
“True AI readiness begins with institutional self-awareness, not technological sophistication.”
This perspective informs Cornell’s AI Leadership for the Public Sector certificate. Across six online courses, participants explore AI fundamentals, data governance, ethical leadership, policy, implementation, and oversight.
They also work with AI solutions, apply frameworks such as the NIST AI Risk Management Framework, and develop governance tools relevant to their organizations.
Measuring AI’s Value Beyond Efficiency
Faster processes and lower costs can make AI investments attractive. But Dr. Austin-Williams emphasizes that government leaders must evaluate results against a broader standard.
“An AI solution can make organizations faster without making it better.”
Consider a system that reduces the time required to process a public service request. Speed matters, but leaders must also examine whether decisions are accurate, services are accessible, and the public’s experience has improved.
“Efficiency alone is not evidence of public value.”
Dr. Austin-Williams encourages leaders to consider decision quality, citizen experiences, workforce effectiveness, and public trust alongside operational performance.
That evaluation must continue after implementation as technologies, institutional needs, and expectations evolve.
Preparing Leaders to Make Responsible AI Decisions
Dr. Austin-Williams developed the certificate for executives, policymakers, program managers, and emerging leaders responsible for decisions about AI adoption and governance.
“This program is for leaders who may never build an AI model, but whose decisions will shape how AI affects society.”
The program connects AI literacy and public policy with practical approaches to governance, helping participants ask better questions, challenge assumptions, and navigate organizational trade-offs.
For Dr. Austin-Williams, the goal extends beyond preparing professionals to implement new technology.
“We’re not simply preparing leaders to use AI. We are preparing them to lead differently because AI exists.”
Lead AI Decisions That Deliver Public Value
Build the skills to evaluate AI, strengthen institutional oversight, and lead responsible implementation in the public sector.
Explore Cornell’s AI Leadership for the Public Sector certificate, featuring curriculum authored by Tasha Austin-Williams and offered online through eCornell.








