AI Fever No Substitute for Sound Governance

AI Fever No Substitute for Sound Governance

Feb 21 , 2026. By Kinfe Yilma (PhD) ( Kinfe Yilma (PhD) - K.Yilma@leeds.ac.uk - is a senior lecturer in law and technology at the School of Law of the University of Leeds. )

In a recent speech before Parliament, Prime Minister Abiy Ahmed (PhD) spoke of an “AI University” to be opened "within months" and a drive to train five million coders as quickly as possible. This vision runs contrary to the government’s earlier, abrupt decision not to open new universities. Recent advances in generative AI have fed a powerful hype cycle. Policymakers increasingly frame AI as a silver bullet for longstanding, complex problems, even as its trajectory and capabilities remain uncertain.

In a speech before Parliament, the Ethiopian Prime Minister announced (or reiterated) the ambitious plan to open an ‘AI University’ within months. A putative AI enthusiast himself, he further flagged the plan to “train five million coders” as quickly as possible.

This is a more recent iteration of the AI hype. In the wake of recent advances in generative AI, the global AI hype is becoming increasingly palpable in Ethiopia. The government has been at the centre of this AI fanfare. Officials tend to portray AI as the silver bullet for longstanding and complex problems. Of course, policymakers have long recognised the potential of information and communication technologies in fuelling developmental ambitions. Successive national ICT policies and digital transformation strategies highlighted the importance of harnessing new and emerging technologies for development.

But the advent of AI appears to be driving them into an unprecedented level of hype. That should be a cause for concern. Such unchecked technoenthusiasm would distract from a clear-eyed view of emerging technologies and their widely recognised risks, and would also divert the government's attention from its principal role in regulating those technologies in the public interest.

The decision to open an AI university is staggering on many levels. The announcement came not long after the government's abrupt and largely unpersuasive decision not to open new universities. It is unclear why it sought to walk back this decision by establishing a higher learning institution dedicated to AI. But one cannot think of a sensible reason.

Many public universities already provide the foundational training needed for AI research and development. In the unlikely event that existing university curricula do not provide state-of-the-art higher education training, the most sensible approach would be to adapt them.

One should also add that the Ethiopian Artificial Intelligence Institute (EAII), a federal agency accountable to the Prime Minister, has a statutory mandate for AI research and training. The Institute is tasked by its establishment legislation, among other things, to “facilitate the development of manpower in the sector in collaboration with domestic and foreign education institutions”. While it has other broader mandates, the Institute is also envisioned as a public institution of AI research and training. This would thus complicate the decision to invest limited and hard-earned public funds in establishing an AI university.

But that is not the only concern. The decision appears to be driven by the assumption that university education is indispensable in making Ethiopia a regional AI hub or to achieving success in AI research, development, and deployment. But none of the countries that have thus far taken meaningful steps in the AI domain has sought to make establishing a dedicated university a priority. In fact, many of the major AI companies based in the US are either founded or staffed largely by technologists with little or no higher-education credentials. There is little evidence that higher education in emerging technologies leads to success in this field.

The logic behind the decision is also unsound. Should one rush to establish a university every time a particular technology emerges, especially when a suite of technologies whose trajectory is not yet fully clear?

Not only is AI’s trajectory, but also its capabilities, yet to be fully grasped. As much of the world is mesmerised over the capabilities (or odd hallucinations) of ChatGPT, the nudification of women by Grok, or the discriminatory outcomes of Amazon’s hiring AI tool, quantum technologies are on the horizon. The logic behind the AI university might, as a result, mean that a dedicated ‘quantum university’ will have to be installed in the next decade or so. That would be a slippery slope of policymaking.

Policy decisions should not be driven by hype and rosy pictures portrayed by actors with a stake in the outcome. Importantly, the government should focus on what it has the mandate, duty, and perhaps the capacity: i.e. governance.

Early deployments of AI across various jurisdictions have exposed a range of risks. From the blatant discriminatory use of AI in policing and employment to infringements of personal privacy, AI tools pose serious risks to the enjoyment of basic human rights and dignity, as well as to the natural environment. To overcome such risks, policymakers have taken a series of governance measures in recent years.

What I call the ‘AI governance rush’ began in 2018 in Europe and has since proliferated globally. African states are slowly turning attention from hype to governance. This took mainly the form of national AI strategies and policies. In the past few years, over a dozen African states have launched such policies. But studies on emergent AI governance initiatives in Africa have shown that the focus remains on the opportunities AI offers rather than the attendant risks.

Ethiopia, too, has taken some governance initiatives. A national AI policy was launched in June 2024, with a sequential AI strategy reportedly underway. As with other African initiatives, much more hope is placed in AI's benefits than in its risks in formative policy initiatives. That is consistent with the overall incautiously optimistic view of AI among policymakers. Nowhere in his speech did the Prime Minister, for instance, mention the risks of AI.

However, the Institute is reportedly developing legislation that is inspired largely by the European Union’s AI Act. And, to a degree, the draft legislation recognises AI risks and places the Institute at the forefront of addressing them. But much remains lacking in AI governance.

The AI bill envisions the Institute as the primary regulator with a broad range of functions. It is envisioned as a centre of AI research and training and as an entity for developing and deploying AI tools necessary for the country's development. It is also tasked with initiating AI policies and legislation as well as regulating the sector. This all-encompassing mandate of the Institute raises several concerns. One relates to the question of institutional independence. An entity accountable to the Prime Minister can hardly discharge its regulatory function autonomously while still being involved in AI development and deployment.

The question of whether the Institute is the right entity to regulate AI remains unaddressed. Indeed, hype and the overall technoptimism that animates the present government appear to have played a role in the Institute's creation. The Institute was established only two years after the creation of the former ‘AI Centre’, which it ultimately replaced and assumed its expansive mandate, including in AI regulation. But under Ethiopian law, institutes are not conceived as regulatory bodies.

According to the law that defines the powers and functions of the executive bodies of the federal government, an institute is defined as “focuses on training, study, research and consulting services”. And such entities would ordinarily be accountable to relevant ministries. A good case in point is the now-defunct Technology & Innovation Institute, which had no regulatory function.

A similar approach should have been followed in establishing the AI Institute, with regulatory functions delegated to other and more appropriate entities. Created amid hype, policymakers charged the AI Institute with extensive mandates that it lacks the capacity and institutional independence to fulfil. Since its establishment, the Institute has been largely opaque, with limited access for stakeholders, including experts in the field.

It is against this backdrop that, the decision to create an AI university should be considered. Channelling resources towards establishing a university rather than putting in place a robust regulatory framework for a technology with serious risks is ill-advised. The effort to support innovation in emerging technologies should not distract one from the urgency of preparing for those risks through meaningful governance arrangements.

Taken together, the policy approach towards emerging technologies should not be driven by hype but informed by local contexts, global realities, and, of course, expertise. More importantly, the government should focus on governance and leave AI development to the private sector. This will be key not only in envisioning a governance approach that supports innovation but also in effectively responding to the risks associated with emerging technologies like AI.

That way, the government can turn its attention and resources towards addressing pressing problems that continue to plague the country.

PUBLISHED ON Feb 21,2026 [ VOL 26 , NO 1347]