Read time: 3 minutes

AI and machine learning – companies urged to brace for impact

By , Portals editor
Africa , 27 Oct 2017

AI and machine learning – companies urged to brace for impact

Elon Musk's warnings about the potential risk of artificial intelligence (AI) are well documented. The co-founder of Tesla and OpenAI is on record as linking this technology to a possible outbreak of a third world war, and his concerns over AI continue to be debated by the likes of Bill Gates and Mark Zuckerberg in the public domain. CNBC online quoted a tweet from Musk which read "AI will be the best or worst thing ever for humanity." It has also been widely reported that he participated in a petition calling for the UN to help enforce a ban on lethal autonomous weapons.

Overall, this reservation over AI resonates with Mark Walker, Associate Vice President for Sub-Saharan Africa at IDC, who also believes the technology warrants forethought and consideration in terms of the impact it will have, both good and bad, on society.

The market research and analysis firm predicts rapid growth for AI, particularly in early-adopter environments such as customer analytics, customer relationship management, supply chain management and security.

It also believes these analytics-heavy environments are a natural fit for machine learning and the competitive advantage gained from AI will quickly evolve into a clear competitive disadvantage for organisations that are slow to deploy these technologies.

As AI and machine learning technologies mature, they will increasingly become the norm within core workloads, according to the IDC.

Says Walker, "Technology has good and bad... AI definitely has a place in the world, I think algorithmic machine-learning-like stuff definitely has a place in the world, but we need to be super aware that it brings both good and bad. We need to start thinking seriously 'what does this mean for the social contract'? That kind of stuff needs to be considered, and built in to this. The problem is that it will be a conflict, because the more social good you build in, the lower your margin – that is the perceived trade-off. Is that a truth is the question."

The reality of Africa's market is that there is progress and development with regards to this technology, specifically around smart cities and the development of related real-time systems, for example.

Walker references the Konza Techno City development in Nairobi as an example of projects or what he calls 'sparks of innovation' that feature across the continent.

Konza Techno City is intended to be a smart city focused on enhanced service delivery, including municipal services.

"People get the fact that yes technology can be wonderful. The reality where you bring it down to the actual 'can this happen now' – it can happen now in very few places in the world and in specific industries and specific use cases," he says. "If you are sitting in California and working with Elon Musk in the Tesla factory, a lot of this stuff that we consider in Africa to be near-term science-fiction, is a reality. However, that doesn't mean it doesn't exist, it does mean that expectations need to be managed."

Walker is not dismissive of the aggressive marketing behind AI and machine learning and says this is important because it raises awareness of the potential value that this technology offers.

And this becomes particularly relevant when it comes to the impact of AI and machine learning on the workplace.

Considering the soft issues

"Human communication is fundamentally changing and if you look at 'what does this mean in the world of work' for instance, the world of work is already changing and that is not just in America at Tesla... it is happening at Volkswagen Port Elizabeth," says Walker.

He visited the manufacturing plant and witnessed the use of robotics on the assembly line. "This company has, I believe, in excess of two hundred robots that assemble Polos."

But when leveraging AI and analytics to bolster operations and ensure efficiency, should businesses take into consideration 'soft issues' like people, for example?

"Yes... because an algorithm is by its nature biased, by who wrote the algorithm. So if I come from a middle class South African family with a certain value system and certain way of thinking, the way I write the algorithm, I will use the stuff I learned at varsity about how to compile flow charts and algorithms, I will use those tools to define the outcomes," says Walker.

As he explains, these outcomes (like making a process more efficient to make more profit) are all good and well, but the way one gets there and how efficient one makes the outcome is pure science and will not factor in social/HR aspects.

So when an algorithm for a company is being written, aspects like total cost of ownership, cost to income ratio and return on investment are all considered. The focus is on profitability and this means cost management and efficiency.

"So when I write that algorithm, those are my inputs and I am going to write the algorithm to make the client happy and I am going to design accordingly," Walker continues.

From an IDC point of view, the expectation is that as AI and machine learning technologies mature, they will increasingly become the norm within core workloads.

According to the company this will correlate with an increased reliance on AI-driven decision support within businesses and will drive significant commercial advantages. However, there is also the risk of making Carew CC-type mistakes. This is a risk that businesses should take steps to mitigate by establishing protocols for embracing AI, along with a formal approach to analysing the potential social and reputational impact of AI-driven business decisions.

Daily newsletter