The adoption of Artificial Intelligence (AI) by companies must be accompanied by control, governance, and risk management mechanisms to ensure that efficiency gains do not introduce security vulnerabilities. That was the position advanced by Ernst & Young (EY) during a session titled 'Artificial Intelligence, Risk and Decision-Making: Executive Decision Simulation in the Face of AI Risks', held on Thursday morning in Maputo, which brought together representatives from various companies to discuss the challenges and opportunities associated with the technology.
For Shamein Popat, EY Manager for Digital Assurance & Technology, AI has moved beyond being a purely technological concern and has become integral to business decision-making, risk management, and trust. In her view, organisations must therefore change the way they approach the technology, avoiding the trap of limiting its use to the automation of existing processes.
'Artificial Intelligence is not only for doing the same things faster. It is also for rethinking the problem,' she explained.
Among the principal concerns identified during the session were data security and efficiency, the reliability and accuracy of AI tools, and a lack of specialised knowledge and skills within organisations.
Popat also cautioned against growing dependency on these tools and the risk of users accepting AI-generated outputs without verification.
'We cannot blindly trust the tool. It is important that we review what the tool produces,' she said, referring also to so-called AI 'hallucinations' — instances where systems generate incorrect responses or fabricated information.
For António Esteves, EY Senior Consultant for Assurance Technology Risk, one of the key challenges facing organisations is scaling pilot experiments into solutions that are genuinely integrated into business operations.
According to the specialist, many companies are already experimenting with AI tools but encounter difficulties moving to full deployment. The main obstacles include unclear governance structures, the absence of defined accountability, and difficulty in measuring the value generated by AI solutions.
The need to strengthen security in AI use was also acknowledged by session participants. Aquiliama Amade, Head of Cybersecurity at Moza Banco, described the session as relevant for addressing a technology already in use by some individuals and companies, while others remain at an early stage — largely due to concerns around information confidentiality.
'Participating in this session gave me a different perspective on the use of Artificial Intelligence. I came to understand that we must use it responsibly and consciously in order to mitigate the various risks it can bring,' Amade said.
Denis Zulo, Head of Risk at BCI, said the technology has the potential to simplify tasks and streamline processes, and could prove useful in his role at the institution. He argued that knowledge of these tools should be more widely disseminated among companies and organisations.
Source: Diário Económico
Original article: https://www.diarioeconomico.co.mz/2026/09/10/trends/ai/ernst-young-defende-maior-controlo-e-governacao-na-adopcao-da-inteligencia-artificial/
