Tuesday, 22 September 2026
Artificial intelligence is forcing universities to make many of the same decisions now confronting large companies: how quickly to adopt new technology, how to protect sensitive information, which processes should be automated, how employees should be retrained and how much authority should remain with people. At Ai4 2026 in Las Vegas, the presidents and chancellors of three historically Black universities argued that becoming an AI-ready institution is therefore not primarily a technology project. It is a leadership and organisational transformation challenge. The discussion brought together Melva K. Wallace, president and CEO of Huston-Tillotson University in Austin, David K. Wilson, president of Morgan State University in Baltimore, and Karrie G. Dixon, chancellor of North Carolina Central University in Durham. Their institutions differ in size and research profile, but all three are trying to incorporate AI across teaching, research and administration rather than confining it to computer-science departments.
The central difficulty, Wallace argued, is that universities are preparing students for a labour market that employers themselves cannot yet define. Businesses increasingly want graduates who are comfortable working with AI, yet many companies are still determining how the technology will change their own organisations. Universities therefore face two transformations simultaneously: they must prepare students for an AI-enabled economy while also redesigning their own operations around the same technology. Universities employ thousands of people across finance, human resources, student services, admissions, facilities, research administration and security, and those activities contain many of the same repetitive and information-heavy processes found in corporations. Dixon described North Carolina Central’s effort to examine where AI could improve administrative efficiency, including recruitment and the processing of large volumes of job applications, while ensuring that adoption is guided by institutional strategy rather than introduced piecemeal by individual departments.
The lesson for other organisations is that AI adoption cannot simply be delegated to an IT department. University leaders on the panel repeatedly described AI as something requiring executive involvement because decisions about automation affect employment, governance, culture, cybersecurity and the organisation’s mission. A president or CEO does not need to understand every technical detail of a model, but does need to determine where the technology creates value, what risks the institution is prepared to accept and which responsibilities cannot be delegated to machines. Morgan State provides one of the clearest examples of this approach through Obsidian, its internal sovereign AI platform designed to provide secure access to institutional knowledge while maintaining greater control over university data. Rather than automatically sending sensitive university information into external public models, Morgan is building an environment intended to retain stronger control over data, security and how AI is used across the institution.
That approach mirrors an emerging corporate debate around sovereign and private AI. Businesses increasingly want the capabilities of large models without allowing confidential information, customer records or proprietary knowledge to move freely through third-party services. Universities face particularly sensitive versions of the same problem because their systems contain academic records, financial information, research data and personally identifiable student information. Morgan is also connecting this infrastructure strategy with academic development by integrating AI more widely into teaching and institutional operations. Wilson’s broader argument was that universities should not simply consume technology developed elsewhere but should build enough internal capability to shape how AI operates within their own organisations and give students practical experience in the process.
North Carolina Central University is pursuing a similarly broad strategy, centred on the idea that AI should not become concentrated within a small number of technical departments or students. Dixon argued that if some faculties, employees and students become highly capable AI users while others remain excluded, universities risk creating a new internal divide between those able to benefit from the technology and those left behind. The university’s AI institute, new facilities and expanding academic programmes are therefore part of a wider attempt to make AI literacy an institution-wide capability involving academic departments, administration, information technology and security. The objective is not merely to produce more AI specialists, but to ensure students in different disciplines understand how the technology will affect the work they eventually enter.
Huston-Tillotson is approaching the challenge from another direction: partnerships and workforce development. Based in Austin, one of America’s largest technology centres, the university has been building relationships with technology companies while positioning AI skills as part of wider career preparation. Its HBCU AI Conference and Training Summit brings universities, students, companies and policymakers together around responsible AI adoption, workforce readiness and applied training. For Wallace, partnerships are important partly because the technology is changing too quickly for universities to develop expertise entirely internally. Companies can provide access to platforms, training and employment pathways, while universities contribute research capacity, students, faculty expertise and an understanding of how emerging technologies affect communities.
Yet the panel repeatedly returned to a problem that cannot be solved simply by acquiring more technology: trust. Dixon described technology as advancing faster than institutions can build confidence around it. Faculty members may worry about academic integrity, employees may fear job displacement, students may question whether learning still matters when AI can generate answers instantly, while administrators must determine which systems can be trusted with sensitive data. That makes change management as important as technology deployment. Employees who have performed a process the same way for years cannot simply be given an AI system and told to use it. They need training, clear expectations and an understanding of why the workflow is changing. University leaders therefore face essentially the same challenge as executives introducing AI into banks, property companies, manufacturers or professional-services businesses: successful adoption depends heavily on whether people believe the technology will support their work rather than simply eliminate their roles.
Cybersecurity was another major concern. Wallace identified protection of student information and the growing threat of ransomware as among the issues that concern university leaders most. Higher-education institutions hold large volumes of sensitive information while operating complex networks used by students, researchers, faculty and outside partners. Introducing additional AI tools can expand that technological environment further, increasing the importance of identity management, data controls and careful vendor selection. At the same time, Wilson raised a different risk: that excessive reliance on AI could weaken intellectual development. Universities exist partly to force students through the difficult process of learning, analysing information and developing judgement. If AI removes too much of that intellectual effort, institutions could produce graduates who know how to obtain answers but are less capable of determining whether those answers make sense.
That concern may be just as relevant to companies. AI can reduce the amount of junior analytical work required in fields such as finance, law, consulting, journalism and real estate, but those activities have traditionally also been how younger employees develop professional judgement. If organisations automate too much entry-level work without replacing the learning embedded in those tasks, they could eventually find themselves without experienced people capable of supervising the technology. The presidents therefore placed particular emphasis on critical thinking, adaptability and judgement. The objective is not simply to prepare students for a specific set of AI tools, because those tools could be outdated within a few years. Universities instead need graduates who can learn continuously, understand new technologies quickly and decide where they should or should not be used.
That represents a significant change in how workforce preparation is understood. Universities have traditionally designed programmes around identifiable occupations, while AI makes the future occupational structure less predictable. Institutions may increasingly need to prepare students for roles that do not yet have established job titles. Technical literacy remains important, but so do communication, creativity, teamwork, conflict resolution and the ability to make decisions when information is incomplete. Wallace emphasised the continuing importance of what are often called soft skills, arguing that AI does not remove the need for people who can communicate, collaborate, adapt and behave professionally.
For commercial real estate and other traditional industries, this has direct implications. Companies may not simply need more data scientists. They will increasingly need property professionals who understand leasing, investment, construction, asset management or building operations while also understanding how AI can be inserted into those workflows. Universities capable of combining domain expertise with AI literacy could therefore become important suppliers of the next generation of professional talent. There is also a physical investment dimension. AI strategies in higher education increasingly require specialised laboratories, collaborative teaching environments, computing infrastructure and research facilities, showing that university AI adoption is not entirely virtual but is beginning to influence capital spending and campus development.
The wider lesson from the Ai4 panel is that institutions should avoid treating AI as another software upgrade. The technology cuts across workforce strategy, organisational culture, information security, infrastructure, education and governance. Introducing it therefore requires leadership capable of connecting those separate issues to the institution’s wider mission. Universities may be unusually useful places to observe this transition because they must confront both sides of the AI economy simultaneously: they are employers trying to become more productive organisations, but they are also responsible for educating the people who will work in the transformed economy.
Their experience suggests that the organisations most likely to benefit from AI may not necessarily be those that adopt it fastest. They may be those that combine technological ambition with sufficient attention to data, trust, skills and human judgement. The central challenge is therefore not simply how much AI an institution can deploy, but whether it can introduce the technology without weakening the qualities it ultimately depends upon people to provide.
Source: CIJ.World Research & Analysis Team
