Sam George Refutes Claims He Tasked KNUST to Build Twi AI Model

Communications Minister Sam Nartey George has dismissed reports suggesting he instructed the Kwame Nkrumah University of Science and Technology (KNUST) to develop a Twi artificial intelligence model.

In a Facebook post on October 12, 2026, the minister said a JoyNews caption misrepresented his engagement with postgraduate researchers at KNUST’s Responsible AI Lab.

“I never asked anyone for a Twi AI language model, as reported,” he stated, sharing a video of the discussion to clarify what transpired.

The clarification followed a JoyNews report titled “Sam George tasks KNUST to develop Twi AI model as Ghana plans local‑language artificial intelligence.” The report described the minister as urging KNUST experts to lead the development of a Twi AI system.

However, Sam George said the caption was misleading and did not accurately reflect his conversation with the researchers. While acknowledging the media’s role in public service, he stressed the need for factual and unambiguous reporting. His Facebook post also tagged the JoyNews page.

What the Video Actually Shows

The video features a postgraduate researcher explaining his work on code‑switching and speech recognition—a common linguistic practice in Ghana where speakers mix English with local languages within a single conversation or sentence.

He outlined challenges speech‑to‑speech systems face when processing multilingual or mixed‑language input and discussed his doctoral research aimed at improving AI models’ ability to understand such speech patterns.

During the engagement, the minister asked questions about existing Ghanaian language datasets, ongoing academic projects, and opportunities for collaboration. He referenced work by researchers at the University of Ghana, Google, and other institutions developing resources for local languages.

The conversation therefore centered on ongoing research and language‑technology development, not a directive from the minister instructing KNUST to build a Twi AI model.

Broader Context: Local‑Language AI Still Matters

Although Sam George disputed the report’s portrayal of his engagement, he did not dismiss the importance of developing AI systems for Ghanaian languages.

AI models trained primarily on widely used international languages often struggle with local languages, accents, and everyday speech patterns. This limits their usefulness for people who communicate mainly in indigenous languages.

Speech‑based AI could help bridge this gap. For example:

• Farmers could receive weather, crop‑disease, and extension information in familiar languages.
• Patients could access healthcare guidance more easily.
• Public‑service information could reach communities with limited literacy.

These applications depend on robust datasets, effective models, and sustained research.

The video also underscores the importance of code‑switching research. Since many Ghanaians naturally blend English with local languages, AI systems that understand mixed‑language speech could be more practical and widely adopted.

Source: Louis Awortwe