Use it to build teams, but never to expect it to replace humans.
On the 16th floor of the 180 Lounger, undoubtedly the only building in Cape Town with the best and uninterrupted views of the city, IT’s leading minds gathered for an enlightening and informative discussion on artificial intelligence (AI) and the CIO at CIO SA’s Summit.
AI is one of the most spectacular and talked-about innovations of our day; in a personal blog post, Bill Gates referred to artificial intelligence as “as revolutionary as mobile phones and the Internet.”
The summit explored the concept of the chief artificial intelligence officer (CAIO), and whether the CIO should be the new CAIO.
The word about the CIO South Africa Community had travelled to Cape Town, and the support was palpable at a filled 180 Lounger. The event would not have been possible without the help of CIO South Africa’s principal partners, EOH and MakwaIT, as well as executive partners BCX and Workday, and associate partner Perpetuuiti.
While the list of opportunities and threats posed by AI was seemingly infinite, CIOs admitted that the main difficulty with AI was that people, particularly the workforce, still saw it as a threat to their jobs rather than something that could help them do it better.
“The biggest challenge I’ve seen is the ’people challenge’, and people being worried about whether AI will replace their jobs. However, the human aspect can never be replaced,” said Hungry Lion head of IT Shalendra Singh.
Chris Shortt, an experienced retail CIO, looked at the use case of AI in the retail arena and said:
“We previously didn’t have the tech for aggregation of data in the retail space. However, the biggest challenge is to try to bring people along. I think we should be using technology to augment people,” Chris said. “Getting people to trust system generated outcomes is another challenge; they might not have the statistical understanding of it, which is why we need to create an appetite for the use thereof.”
Let’s build AI teams of the future, but let’s get it right!
On the other hand, Kosta Kontos, lecturer of the Data Science Leadership course at the Graduate School of Business, University of Cape Town, saw AI as a chance to construct high-performance teams that will work with it rather than against it. He gave some pointers on how CIOs might put together a future-ready AI team.
Kosta believes that data science should be prioritised. Because “even if we talk about putting together an AI team, we’ll still rely on data science.”
If you want to build an AI team of the future, you need to look at where you’re at now. Maybe consider a hybrid approach of insourcing and outsourcing talent.”
On talent, Kosta believes that data science and data engineering is an often overlooked skill, since data scientists clean up the data and make it optimal to use for AI operations.
“Equally so, there are not enough good data engineers in the market and where they are, they get scooped very quickly. You need to be more transparent about remuneration upfront so that they know what’s being offered for that role.”
According to Kosta, the recruitment process also entails spotting misleading marketing. He was alluding to the CVs that people submit when applying for jobs. “We try to demystify the snake oil that is prevalent in this industry, because people often send through heavily embellished CVs that never live up to what is written in them during an interview.”
If you want to attract young people, Kosta has a clever trick. “Create an excellence centre – give it a cool name because this is the type of working environment people want to return to,” he said. “Young data scientists in particular, love to work for organisations where they can learn. In addition, offer courses that can be done during working hours, however, that individual needs to be able to present what they’ve learned. Also consider unlimited textbook budgets, where your people can have unlimited access to educational books.
“Don’t just focus on the areas of success, but also focus on the failures in order to learn from them. In addition, build a meritocracy, where the ideas of the most merit succeed, and try to build an internal forum where teams can share their ideas freely.
“Most importantly, delegate and hire the best people and get out of their way. In addition, avoid the ‘brilliant jerks’ – individuals who are very smart but cannot be integrated or connect with the rest of their team members.”
AI: A look at the past, and a look at the future
Steven Sidley, author, blockchain guru and Professor of Practice at UJ, followed with a quick ‘lecture’ on the history of AI. Steven is well-versed in AI and its capabilities as a former global CIO and the first person in South Africa to take an AI course.
“Artificial intelligence is entirely different from all the transformative technologies that have ever been invented,” he said. “AI in principle learns autonomously without our input, it is growing smarter exponentially, and we have no idea what it will look like many years from now. Nobody saw ChatGPT coming: this thing is coming at us at an accelerated speed.”
Steven believes that it cannot replace humans, but what it produces or gives people alternative options in terms of what they want to consume, whether it be text or music. It has security risks, legal risks (copyright issues), and ethical risks (it being used to cheat or do tasks better), but it’s here – for the long haul – and there is a personal and business case for it.
“Microsoft Copilot, Google Workplace, Bloomberg GPT, Adobe Firefly, Whisper, iGenius, and even a chinese company had an AI CEO for seven months, everyone has entered the race.”
There is also some excellent research being undertaken in the fields of language and language modelling. Jade Abbottt, director at Lelapa and co-founder of Masakhane, and Farayi Kambarami, head of central planning and data at Woolworths Food, provided guests with an insight.
“I wanted to develop technologies for African languages, but I had to weigh the risks. You need a massive pool of data, and AI tools were specifically designed for the English language, but what about the other languages?” he asked. “I’ve also experimented with ChatGPT, specifically with my own native language, Shona, and the results were, to put it mildly, disappointing. There is still opportunity for improvement.”
“Many companies are battling with the mass market conversation. Everyone is trying to understand how to break through in that space, because there isn’t a lot of information that exists about African cultures. We’re employing a lot of European ways of doing things, which is making us irrelevant for a large part of our own markets. You can capture so much more of the African market by just working in some of the other languages on the continent,” said Jade.
“If you move the clock 200 years into the future, and archaeologists try to understand today, they are going to excavate the information digitally. But, based on what exists online right now, most of Africa doesn’t exist. We need to build technology that can enable people to use African languages on the internet and allows us to express our cultures,” Farayi concluded.