Researchers at the Kwame Nkrumah University of Science and Technology, Kumasi (KNUST) have developed an artificial intelligence (AI) system that can detect cyberattacks on factories, power plants and water treatment facilities with more than 99% accuracy, while keeping sensitive operational data private.

The study, published in the Journal of Computer Networks and Communications, presents a new cybersecurity framework that allows industrial facilities to work together to improve their AI systems without sharing confidential data.

Lead author Dr. Emmanuel Osei Owusu said the research addresses one of the biggest cybersecurity challenges of the Fourth Industrial Revolution.

"This study addresses a major challenge of the Fourth Industrial Revolution: how do we make industrial systems smarter and more secure without moving sensitive operational data to one central location?" he said.

He explained that as Ghana's factories, electricity networks, water systems and other critical infrastructure become more connected through digital technologies, they also become more vulnerable to cyberattacks.

"Greater connectivity improves efficiency, but it also creates more opportunities for cyberattacks," he said.

Instead of sending sensitive information to a central computer for analysis, the new system allows each facility to analyse its own data and improve its AI model locally. Only small, compressed updates are shared with other facilities, helping to protect confidential information while reducing internet bandwidth and computing costs.

The system can also identify suspicious or false updates from compromised computers, helping to keep the overall AI model secure.

The researchers tested the framework using internationally recognised industrial cybersecurity datasets. The system detected attacks with more than 99% accuracy while using far less data than conventional methods, making it suitable for industries with limited computing resources.

Dr. Owusu said the findings show that advanced AI-based cybersecurity can be adopted even in countries where computing infrastructure is limited.

"For Ghana, this means advanced AI-based cybersecurity does not necessarily require massive computing infrastructure or continuous transfer of sensitive data," he said.

"Industrial facilities can learn from one another while keeping their operational information private."

He said the study also demonstrates that Ghana can contribute to developing technologies for the Fourth Industrial Revolution rather than simply adopting them.

"Ultimately, this research demonstrates that Ghana cannot only be a consumer of Fourth Industrial Revolution technologies but also a contributor to the technologies that will secure them," he said.

Story: Emmanuel Kwasi Debrah