Kathy Gibson reports – Artificial intelligence (AI) is not just reshaping technology, but is being used to empower people and communities around the world.
Indeed, the technology holds both potential and peril, says Arthur Goldstuck, MD of World Wide Worx.
It’s been 75 years since Alan Turing published his scientific paper envisaging AI. Since then, we have been through two slumps, or AI winters, where the technology failed to meet its potential.
Throughout its history, Goldstuck points out, we haven’t had the computing power to properly drive AI.
“Then, on 30 November 2022, Chat GPT was unleashed on the world – and suddenly everyone in the world was aware of AI.”
But even before then, AI was quietly going mainstream and a number of industries were using it to transform the way they do business. For instance, in 2018, Spanish football authorities were using AI to plan the best times and venues of games, while coaches were using the technology to get the best performance from their players.
As a result, for much of the last decade, Spain dominated the football world. “It’s not the AI would win games, but gave them enough of an edge for them to dominate.”
The key to this success – and for any organisation – is the availability of vast quantities of data together with the ability to process it, Goldstuck says.
The automotive industry is another long-time user of AI, with autonomous vehicles in use for at least three years now.
The lack of trust in these vehicles has slowed their adoption, and Goldstuck points out that trust is the second major need for AI to succeed.
The manufacturing industry is no stranger to the use of robots, with motor manufacturers especially early adopters of the technology.
The lessons from this industry have demonstrated that automation and robotics have the ability to create new jobs, scotching the belief that AI will cost people jobs.
AI in conjunction with edge computing in the form of drones is massively improving agricultural productivity, crop health, combatting pests and measuring soil conditions among other things to radically improve food production.
In financial services, banks and insurance companies use AI to combat fraud and theft, as well as to tailor-make financial products on a person-by-person basis.
“So the third lesson we have learnt about AI is this: if it’s digital, AI can do it,” Goldstuck says. “Thinking is not digital – so if it requires thinking, AI cannot do it. But otherwise, it can.”
Since the advent of generative AI (GenAI), we have seen massive shifts in its use in healthcare, Goldstuck says.
“AI can process vast amounts of data much faster, helping to detect patterns that medical personnel might have missed. This is particularly useful in analysing images and scans, with the ability to drill down into various images, putting it together with information from the patient and doctor to come up with a clear conclusion that the doctor can then consider.
“It means medicine can become truly personal and completely accurate.”
Of course, AI cannot replace the doctor because it cannot deal with unpredictability or ethical dilemmas, Goldstuck adds.
AI is also making massive strides in drug discovery, with its ability to process vast quantities of data.
GenAI is making this possible, with its use of large language models (LLMs) – and there is a vast environment of LLMs producing AI tools that anyone can use, in any language.
“So it is no longer a world dominated by one model, or one LLM. What is transforming AI is the ability to use LLMs to create new content based on existing content.”
In South Africa, recent research from World Wide Worx asked 100 of the largest companies how they are using GenAI: 45% said they were planning to use it; 25% said they were using it unofficially; and just 11% were using it as an officially-approved tool.
“But this shows that 90% of South African companies were using GenAI in one form or another.”
Most are using it for text content creation, and many are planning to use it for video creation too. On the more technical, IT or engineering side of AI, usage falls off significantly, but are on the radar.
The impact of GenAI has been huge – and it has happened quickly.
Ninety-five percent of companies reported a positive impact on productivity. “Yes, AI will kill off a lot of jobs, but it will boost a company’s competitiveness, and thus new jobs will be created.”
Most companies also found a positive impact in competitiveness and turnover as well.
The areas where the impact is seen as negative is HR processes, staff retention and staff recruitment. “So where humans are involved, there is far less of an impact. Where humans are required, AI does it well.”
The big elephant in the room when it comes to AI is ethics. It’s no secret that biases exist in algorithms that lead to discrimination. This is exacerbated because “black box” AI models lack accountability and trust.
In addition, AI is being used for malicious and fraudulent purposes, from misinformation to identity and theft, which poses massive challenges to existing systems.
To counter this, robust ethical frameworks and regulatory oversight is needed in areas like privacy and security.
Education is an area that could benefit hugely from AI – but there are many ways it could have negative impact.
“Given the ethical challenges, should AI be allowed into education?” Goldstuck asks.
Some experts believe it will have a negative impact on education, especially students’ accelerated use of the rapidly-evolving technology.
On the flip side, the World Economic Forum now recognises that AI will have a massive, usually positive, impact on jobs and education – but the education system needs to be overhauled to be able to take advantage of it.
“In South Africa, we have an advantage in our adoption of outcomes based education, which doesn’t rely on student parroting learned content back.
“Where OBE has failed is where teachers did not have the right tools to execute it. AI now gives them to tools to do so, and the possibility to turn OBE into the modern education system it should be.
“AI changes the approach of education from one hole fits all to personalised teaching and assessment,” Goldstuck says.
“And this personalisation extends to all areas of endeavour, helping humans to be more productive, more competitive, and more personalised.”