Technology expert Ali Kingston Mwila has called for stronger legal and ethical frameworks to govern artificial intelligence (AI) in Zambia, warning that the technology could expose individuals to data breaches, discrimination and cyber security risks if left unchecked.
Speaking in an exclusive interview with Zambian Business Times-ZBT Mwila noted that the rapid development of AI systems, particularly large language models capable of generating complex computer code, requires governments to establish clear rules to ensure the technology developed is used responsibly.
He mentioned that although AI systems do not think like human beings, their ability to learn from data and generate complex outputs could create challenges when the systems operate in environments where their actions are difficult to understand or control.
Mwila stated the need for Zambia to strengthen its approach to AI governance by ensuring that laws, regulations and ethical standards kept pace with technological developments. “The discussion around AI is so complex.
There is no one approach or solution to it, but it is a learning process and a continuous way of looking at things,” he said.
He suggests that Zambia should consider developing and strengthening laws and regulations specifically addressing emerging challenges associated with AI, including the misuse of personal information, cybercrime and automated decision-making.
However, Mwila noted that the rapid pace at which AI is developing means that governance frameworks would need to be continuously reviewed. “Protecting personal information was particularly important as AI systems increasingly interacted with databases containing sensitive records, including medical information,” said Mwila.
Mwila warned that when such information was inadequately protected, it could potentially be used to profile individuals, creating risks of discrimination in areas such as immigration and access to services.
He mentioned that AI governance should therefore extend beyond information shared on social media to include personal data collected by public institutions, businesses and healthcare providers.
“It is not only the data about what we do on social media, but it is the data about everything, where the medical records are captured about the patient’s illness and everything, and then it is stored somewhere,” Mwila said.
He added the quality of data fed into AI systems was also important because machine-learning systems relied heavily on data to produce their outputs. Mwila said poorly managed or inaccurate data could contribute to biased outcomes when AI was used to make or support decisions affecting people.
He also raised concerns about the ability of advanced AI systems to generate complex computer code, saying identifying errors or vulnerabilities in such systems could sometimes require significant human effort.
Article by Karen Ngulube
