Ask what I can do for AI ...

Computers have been doing amazing things for 60 to 70 years. In science fiction we have become aware of artificial intelligence in the '80s and so on, and it has been predicted that AI would do miracles. But now with these large language models, it has really become a reality for normal people. So, I think the question that people have is what is the difference between AI and what computers did before AI?

Society needs to get a better understanding of AI. A lot of people have this misconception that AI started in the last two or three years. It has become more publicised lately, and that is why people started becoming more aware of AI. But if you look at the history of AI, the term was coined in the '50s already. It has been around!

A lot of work has been done on the foundations of AI, and applied AI-based technologies. What has made it more popular, as you said, is the advent of GPT models which are linked to large language models, which we sometimes refer to as sensationalised AI. This hype has been created around it because you start to see more applications of AI in mainstream society.

There has also been a lot of steam around AI as a science, primarily supported by how the inherent technologies around and supporting AI have grown. If you look at how our PCs have evolved over the last 20-30 years in terms of computing power. Computing power has become more accessible. If you take smartphones – their capacity versus what a PC was 20 years ago - the smartphone is more powerful from a computational storage point of view. It means you can do a lot more at a lower cost, which is creating opportunity for greater adoption of AI-based solutions.

The automation of daily tasks is where the opportunity lies with AI. If you even go back to the beginnings of the calculator, the simple calculator when it was introduced also created a lot of debates as to what would be the future of computations, but the whole element of bringing such technologies was to see where the mundane kind of work can be replaced by an automated approach.

I used to have a small business, renovating classical modern future. We had seven employees – the admin side and the logistics such as deliveries, and then we would have two upholsterers and two woodworkers. How could AI help with the mundane tasks in such a set-up? Is there a type of audit that one can do?

We are already seeing a lot of adoption in mainstream industries. So, if you look at the financial sector, that's one area where AI is becoming very prominent – they are really adopting the use of AI for looking at how you can process extremely large amounts of data in a more efficient manner. What you have is the ability to collect data about an environment, in this case the financial market and connect tendencies, and behaviours, and use that to help make better decisions. What you target is to overcome the human's limitation through the adoption of such a technology. The human is constrained in their ability to process beyond a certain amount of information, and this is when technology has the opportunity to kick it and to overcome.

For example, we do a lot of work in healthcare sites, for instance in cancer detection and therapies by using AI-based technologies to sift through scans. The technology can do work that is beyond the capability of even a specialist in terms of detection, you know, in terms of their ability to pinpoint the exact location of the start of a cancer in our cell formation. In certain cells, you are not limited by time when you introduce a technology like that when a human would have limitations on how many hours they can constantly process. A computer with AI can do unlimited amounts of time.

Beyond that, also to do connections where you have multiple sources of data. For example, one of the projects that we are doing is where you do these scans of tumours and then connect it with things like the history of the individual, DNA samples, and see how the formation and your medical history connect up to find the most suitable treatment that we can propose that can target specific types of cancer. It’s the limitation of the human that brings about this opportunity.

When you bring it down to small businesses, how can we make use of such technologies? Process automation is one area that such technologies can bring about opportunity, so it's to look at your current business environment and look at how your processes are currently configured and how you can then bring such a technology to recommend what are the optimisation strategies you could adopt within your small business to bring about better efficiencies and operational improvement, whether it’s from a cost, a market or a time point of view.  

Okay, so let me give you a specific example. In my online business I used the software Shopify, to host my online shop, and I also used Sage Cloud Accounting, which are two very general and well-known programmes. What I needed was to create a workflow between these two things. For instance, when a sale happened on Shopify, for the profile of the customer to also be automatically created in Sage, and vice versa. That would save a lot of time and manual input. You can find plug-in programmes to do this, but they are prohibitively expensive for small businesses. Here’s my question: Can I conceivably ask an AI system to write a little programme for me that can work, or do I still need a specialist to do it?

Yes, so look, from a generic application point of view, development is becoming more feasible. Just to answer your question directly. If you were to prompt one of these GPT platforms, it is quite easy today to develop fully functional applications by just prompting, and you know indicating exactly the kind of features you would want from such an application. It is quite possible with the kind of tools you have available in the market today. The challenge will come when you are dealing with proprietary software, like Sage, for example, is a proprietary platform. To have plug-ins working with such a platform sometimes might bring about constraints.

But in terms of creating tools purely generated using AI is very possible. I mean, one of my colleagues is doing a basic course to teach students how to create useful mobile apps, and all the code is generated by prompting AI.

That is one part. The second part where there’s a lot of emergences of tools, is the space of agentic AI. AI agents is an area where there’s a lot of development that’s taking place where you create agents to take care of different business processes, and these agents are built to take care of… like you mentioned one element speaking to the other element – each is then built into an agent, and these agents collectively work together. So, the space of agentic AI is also growing quite rapidly, and if you go on the internet today, it is quite easy to use AI to generate these agentic AI’s where you have all these agents for each process, working independently but towards a common goal. It is easy to roll out and very possible today.  

Is it necessary for normal people dealing with these GPT-type systems to learn how to ask better questions to get better answers?

If you are generally computer-oriented, it's quite easy to learn how to get your GPT platform to respond in the way that you are looking for.

If you want very specific types of outcomes, then some sort of training is necessary, like prompt engineering. It's an area that we are looking into to develop a course purely focused on prompting which is the task of how do you prompt in the correct way to get the kind of outcomes that are more precise.

On the basic level of automating daily tasks, what are the most promising areas of development now?

In South Africa, the automation of chatbots as first-responding AI-based platforms to assist clients, and the automation of call centres is large. Also, if you look at the financial sector, we are quite advanced.

Which jobs are most likely to be made redundant by AI developments?

All mainstream careers are currently in my opinion under threat because you can basically get an AI-based solution to start influencing most mainstream careers, so take the medical, the manufacturing, the farming and financial sector, you could basically go across where an AI-based solution could even threaten even advanced careers in the next 10 to 20 years. I mean, take the medical sector. Certain procedures are slowly moving towards being replaced by robotic platforms. They are more accurate, less susceptible to error.

The manufacturing sector, BMW in the US, has their honour test phase of adoption of this humanoid platform to take over some of the manufacturing processes. So, I mean, if you're looking across society, there is a lot of examples where technology platforms that are AI-based are starting to influence, so that is the reality.

Look at the global geopolitical situation between the US and China and the emergence of platforms, I mean, if you look at DeepSeek, for example, being more accessible, it's making the technologies more feasible and adoptable and easier to build solutions, so the rate at which you're going to get more solutions to replace activities in mainstream society, it’s going to become more of a reality.

So how do the youngsters respond to that?

The answer to a great extent is to become more familiar, because there are really just technologies. If you look at technology deterioration, they will need maintenance and this maintenance will require human intervention.

So, getting a career that is oriented to the AI space will become more fundamental? I mean, there was a time that we used to drive that when students coming to a university environment, they need to do computer skills. That will change, it will become more important to become AI-literate as opposed to computer-literate so that that becomes part of your daily knowledge in terms of your skill level. A skill that you need to make part of your package of skills that you are able to demonstrate and mainly because then there will be this need to utilise that skill for intervention for any AI-based solution that is being rolled out.

♦ VWB ♦

BE PART OF THE CONVERSATION: Go to the bottom of this page to share your opinion. We look forward to hearing from you.