The author shares impressions from visiting their rural hometown in Limpopo, where interactions with friends from their roots differ from those with urban residents. Recently, the topic of artificial intelligence (AI) has become very relevant, causing people to worry about how this technology will affect their lives, for example, whether it will lead to job losses.
The questions the author receives boil down to two main aspects: whether AI can help people find jobs and how it will facilitate the use of public services. The author notes that explaining technological concepts in simple language, especially to those concerned about securing food, is a difficult task. He emphasizes that citizens must participate in decision-making processes that affect their lives.
Many official statements about AI swing between excessive promises and strong fears. Some claim that AI will solve all problems, while others create the impression that people have no influence on the future. Instead of informing people how AI can help them, where mistakes might occur, and how to make it useful, the discussion often devolves into fear-mongering.
This approach only increases anxiety about AI, causing people to miss even the most basic information about the tool. They do not know that AI is software trained on examples to perform specific tasks. They are also unaware that AI can help write letters, translate words, sort records, or detect something unusual. For instance, small businesses can use AI for inventory planning, and a nurse can use it to reduce paperwork and increase time spent with patients.
There are also drawbacks that need to be known. Some AI tools can provide answers that sound convincing but contain fabricated facts. They may misinterpret questions or struggle with local languages, as well as reproduce unfair ideas from training materials. This is comparable to how a person confidently giving directions might lead to a dead end if they don't know the road; AI is capable of something similar, so important answers always require verification.
Risks increase when decisions concern health, income, or a person's access to services. Furthermore, there is the problem of incorrect data. If an AI system uses records, for example, about municipal bill payments, without proper verification, it might send an incorrect bill faster, leaving the family with the same problem but accelerating the error. Accurate records, clear division of responsibilities, and a willingness to correct mistakes are required; buying a smart system does not replace this human work.
Leaders are obliged to maintain record accuracy, respond to complaints, and fix shortcomings. An automated response like 'Your inquiry is important to us' is meaningless if no one provides assistance. Citizens expect a reply, a fair explanation, and the ability to correct an error; the phrase 'The system failed' should never end the dialogue, as someone must be held accountable.
Jobs also require an honest approach. AI can transform workflows and reduce the need for certain tasks, but it can also improve people's work and open up new opportunities. The population needs to be prepared for these changes. While benefits for everyone cannot be guaranteed, decisions can be made that support workers and create new chances. Employers must inform employees about changes in advance and involve them in the process. Government, businesses, and universities must promote the acquisition of useful skills, and training should give people more chances to earn money, improve their work, or develop their business.
Accessibility is equally important. A service may function perfectly on an expensive smartphone but be useless to someone who cannot afford the internet. If the system does not understand the language the person speaks, it creates an additional barrier. Help must be available to both a grandmother in the village and a young applicant in a settlement, and services must remain open to those who need help in person. Ensuring personal information protection is also vital; before entering identification numbers, medical records, or private documents into an AI tool, responsible parties must know who and how can access them.
For all decision-makers in the field of AI, three questions should guide any major proposal: what problem will it solve, how will we know it works, and who will answer if something goes wrong. It must start with a clear need, test the idea on a small scale, listen to users, and verify results and full costs before scaling up. Sometimes a simpler change can solve the problem better.
These discussions should take place not only at conferences but also in community halls, villages, schools, and workplaces. People who use public services daily often know where problems arise, and their experience should influence solution development. They should be invited to test new services and provide feedback, as a plan that looks good in a meeting might feel different to the person using it.
The author hopes that conversations about AI will focus on whether it brings visible improvements to people's lives: shorter queues, correct billing, strengthening small businesses, or more time for patient care. This is the goal to strive for. South Africa should explore AI with optimism, asking questions without fear, and assessing progress by how people's lives improve. Good leadership, honest work, and caring for others remain indispensable, as no machine can take on these responsibilities.
