By Philip Antoh
A high-profile partnership between Ghana’s Ministry of Education and tech giant Google, aimed at developing an artificial intelligence tool for schools, has backfired spectacularly, exposing deep-seated tensions over language, regional equity, and the opaque nature of government-tech deals.
The project, announced as a leap forward for digital learning, is now mired in controversy after the exclusion of the Ga and Adangbe languages from its pilot phase sparked accusations of marginalisation against communities in the capital region itself.
In a defensive statement, the Ministry’s Press Secretary, Hashmin Mohammed, insisted the selection of four unspecified local languages was “strictly on a pilot basis,” with a promise to later include all twelve government-sponsored languages.
“At no point is the pilot intended to marginalise or permanently exclude any linguistic or cultural group,” the statement read. But for critics, the damage is already done, revealing a troubling lack of foresight and consultation in a project of national significance.
Behind the technocratic language of “testing efficacy” lies a potent political question: who decides which languages are first in line for digital transformation? The Ministry has refused to disclose the selection criteria or name the four chosen languages, a lack of transparency that fuels suspicion.
Analysts suggest the choices likely reflect a combination of existing digital language data, the influence of powerful regional constituencies, and logistical ease factors that inherently disadvantage smaller or less politically connected groups.
“This isn’t just a test phase; it’s a power map,” argues Dr. Naa A. A. Mensah, a political sociologist specialising in education policy. “The communities in the pilot will shape the tool’s development, creating a lasting advantage. To exclude Ga, spoken at the seat of government, is not just an oversight; it’s a profound political signal that calls into question the entire equitable design of the programme.”
The controversy also casts a spotlight on the role of Google, whose influence in shaping digital infrastructure across the Global South is under increasing scrutiny.
The company’s interest likely lies in acquiring valuable, structured datasets in African languages to train its AI models. The pilot phase essentially allows Ghana to subsidise the creation of this data for a select few languages first.
“The state is providing the legitimacy and the access, while Google gains proprietary data,” notes a tech policy advisor who has worked with the government, speaking on condition of anonymity.
“The question is what guarantees Ghana has that the data for all twelve languages will be developed with equal rigour, especially if the initial pilot is seen as flawed or exclusionary. The risk is we end up with a ‘Google-approved’ linguistic hierarchy.”
For Education Minister Haruna Iddrisu, this is a significant early test. The project is a flagship initiative meant to showcase a modern, inclusive vision. The backlash from the Ga and Adangbe communities—traditionally influential in Accra—represents a tangible political threat and a public relations failure. The ministry’s belated, generic assurances have done little to calm the situation, pointing to a disconnect between its centralised planning and grassroots realities.
The incident follows a pattern where grand technological solutions, often engineered in partnership with foreign corporations, are rolled out without adequate stakeholder mapping or community buy-in. The result is a predictable cycle of announcement, backlash, and damage control.
The stakes extend beyond this single project. Ghana’s attempt to integrate AI into its education system is being closely watched across the continent. If it is perceived as reinforcing existing ethnic and regional fissures, it could serve as a cautionary tale, undermining trust in similar initiatives elsewhere.
The Ministry now faces a choice: it can continue with a rigid pilot plan that has already alienated key stakeholders, or it can pause, engage in genuine consultation, and redesign the rollout to ensure its rhetoric of inclusion matches its actions. The path it chooses will determine whether this AI tool becomes a bridge to the future or a new wall dividing the nation.