Perspective on responsible AI governance, institutional readiness, digital equity, and the controls that make safeguards meaningful in practice.
Doctor of Ministry in Digital Stewardship · May 2026
Scholar-practitioner
Research that becomes practice
Our frameworks began as doctoral research. Dr. Alexis S. Smith earned her Doctor of Ministry in Digital Stewardship, with a concentration in AI Ethics and Governance, in May 2026. That work became the Ascend AI NOW™ Framework for Responsible AI and the AI Power Gap™ Readiness Index that power our client engagements.
Book chapter · Forthcoming May 2027“Digital Stewardship: A Black Faith-Rooted Ethic for Artificial Intelligence (AI) Governance in K-12 Education,” in Black Lives and AI: Innovation, Resistance, and Possibilities for Flourishing, edited by Dr. Randon R. Taylor and Krystal Jackson, M.S. (Palgrave Macmillan / Springer Nature).
Book · ForthcomingClosing the AI Power Gap™, Ascend AI NOW Publishing.
AI Power Gap™ Intel
Reviewed Does Not Equal Governed
A loan officer receives an AI recommendation to decline a caterer who brought seven years of business records and a signed contract. A checkbox says a human reviewed it. The governance question is whether anyone can prove that review was real: what the model used, what the reviewer read, and why the decision stood or changed.
A signature is an event. Judgment is evidence.
Log what the model used and what it ignored, so reviewers verify inputs before outputs.
Require a written rationale in the reviewer’s own words, not a pick-list.
Track override and escalation rates. A review that never disagrees may be a rubber stamp.
Tie every adverse decision to specific, documented reasons a customer can understand.
Assign final accountability to leadership and process, not to the model or a single officer.
Crosswalk: EU AI Act Article 14 (human oversight, including awareness of automation bias and the ability to override) and Annex III, point 5(b) (creditworthiness of natural persons as high-risk); Regulation B, 12 C.F.R. § 1002.9 (specific reasons for adverse action); NIST AI RMF 1.0, Govern function. Directional guidance, not legal advice.
AI Power Gap™ Intel
Acknowledged Does Not Equal Controlled
Governance requires more than risk acknowledgment. Recognizing a risk is an event; control requires capabilities that can detect, contain, monitor, and stop an unsafe process when needed.
Acknowledgment is an event. Control is an outcome.
Detection and escalation are not equivalent to stopping an unsafe process.
Evaluate containment and termination capabilities before deployment.
Set monitoring severity to match the impact and operating context.
Map environment coverage, including relevant integrations and operating modes.
Require evidence that safeguards work in practice, including when systems fail or behave unexpectedly.
Digital equity & policy advocacy
Connectivity, access, and stewardship
K–12 infrastructure and access
Dr. Smith's K–12 technology leadership experience spans infrastructure, connectivity, digital access, and educational policy advocacy. Her roles include CoSN Thought Partner and K–12 CTO Academy Mentor.
E-Rate and public policy
Her E-Rate advocacy and educational materials address the FCC's 2026 review of the program, with CoSN attribution preserved.
What we are building
Business optimization conversation design
We have been building conversation-design prototypes for business optimization discovery that guide leaders from current-state pain points to prioritized AI opportunities.
