AI has the potential to address inequality in South Africa and ensure the country remains connected to the global digital economy. But, this will depend on access to and the ethical use of this emerging technology.
This was the consensus among technology professionals who participated in a panel discussion at the Rectron Summit 2024, the tech distributor’s flagship channel event held in Johannesburg on 12 September
Panelists acknowledged the role AI can play in driving efficiency, improving productivity, and generating new revenue streams, especially in use cases across agriculture, manufacturing, education, healthcare and mining. However, this cannot happen independently of ethical practice.
“There are definitely new opportunities with AI, but it has to be treated with respect,” said Deveshane Naicker, channel lead for Africa at Zebra Technologies.
Model collapse
In his keynote, futurist Dr Craig Wing discussed the impact of model collapse, an emerging trend where machine learning models weaken over time because they are fuelled with erroneous data from other models, including prior versions of themselves.
Forbes online defines model collapse as “what happens when AI models are trained on data that includes content generated by earlier versions of themselves.
“Over time, this recursive process causes the models to drift further away from the original data distribution, losing the ability to accurately represent the world as it really is. Instead of improving, the AI starts to make mistakes that compound over generations, leading to outputs that are increasingly distorted and unreliable,” Forbes states.
According to Dr Wing this trend is fuelling bias and hallucinations within the application of AI.
Rectron’s message to channel
Rectron’s core message to the channel is that AI is not there to replace humans, it’s there to enhance and speed up work, said Jaco Oosthuizen, category head for AI, mobility, and trends at Rectron.
“It is easier to complete a project that is 80% done than it is to start the project from scratch. I think sometimes in the environments in which we are working, it is hard for people to always be innovative… I think AI really helps with that. It will free up time for individuals and give them the opportunity to become more innovative.”
Oosthuizen advised business leaders against making iron-clad AI management decisions before fully testing the technology.
“I think AI is going to go through a journey where guys are going to test it in the ChatGPT format, in the co-pilot format, they are going to test it at executive level, maybe one or two layers down and get an idea of what the risks are, and define the business process that comes with that – and then I think the next step will be the fast-paced adoption.”
Spencer Chen, Rectron's chief executive, said, “Artificial intelligence has been a part of humanity for decades already, gradually progressing from experimental settings in the 1950s to the emergence of machine learning in the 1990s, to now becoming a mainstream reality across people’s personal and professional lives. Today, we are seeing a rapid acceleration and proliferation of AI applications, from virtual assistants, image (and speech) recognition, natural language processing, predictive analytics, autonomous vehicles, health and biotech, and countless sectors embracing the efficiencies and innovative solutions AI can unlock."
While promising, the use of AI raises both ethical and technical questions that need to be addressed, Chen added. These range from the impact on public opinion and jobs to concerns about AI autonomy and the security of AI systems against cyber threats or unintended system activity.
Monetising AI
During a second panel discussion on the ethics of AI in business, AI expert Johan Steyn noted that businesses, under pressure to meet margins and targets, might be tempted to use AI without sufficient regard for ethics or without adequate safeguards in place.
He posed a question to the panel: “Is it possible to use AI ethically and still make money?”
Callan Abrahams, principal AI consultant, iOCO, said the issue speaks to the culture of a business. “Leadership does not trump AI … you cannot fix business management with AI.”
Oosthuizen said ethics is being built into AI and it is important to ‘go deeper’ with the technology, collaborate as much as possible to align ethical considerations with the objectives for use of AI.
Wing agreed, adding that it's essential for companies to explain to board members that using AI without ethical considerations could lead to real, devastating consequences.
Balancing ethics with profitability requires a shift in mindset, Wing continued. Decision-makers must view the ethical use of AI not as a ‘grudge purchase,’ but as a business imperative.
The panelists agreed that as South Africa moves towards a data- and digital-centric fifth industrial revolution, AI will play a significant role. However, this progress cannot come at the expense of ethics and humanity.
Share