No slave to the machine: AI to enable humans

“A fundamental risk to the existence of human civilisation,” says Elon Musk. “Potentially the worst event in the history of our civilisation,” says Stephen Hawking.

DESPITE the reservations of several leading technologists and scientists – most notably among them Messrs Musk and Hawking – Artificial Intelligence (AI) is one of the most exciting and possibly useful technological developments in our history. It has the potential of driving productivity and economic growth by facilitating unprecedented human-machine collaboration.

There are hopes that AI will improve competitiveness and ensure citizen wellbeing. It will liberate us from mundane, unfulfilling tasks and allow us to do things that really matter.

There are currently two prevailing beliefs about the path humankind will take in adopting AI: the low road is one of dystopia, where at worst humankind become slaves to their robot overlords, or at best only a small handful of ultra-wealthy individuals derive benefit from AI. The high road – and this is one I believe we can realistically strive for – is one where AI and associated technologies are used to realise an exponential future that ultimately leads to a more prosperous, just and equal society for all of Earth’s citizens.

PwC estimates that AI could contribute as much as $16 trillion (R188trln) to the global economy and boost GDP growth by 26 percent in 2030. Accenture claims AI could boost productivity by up to 40 percent while enabling the development of entirely new products and services that drive new revenue streams and create new markets.

But in order to take the high road, urgent dialogue is needed to ensure AI solutions are built and implemented with broader societal benefits in mind.

Let’s get one thing out of the way: AI is certainly going to spell the end for millions of jobs. No industry is safe. Some industries will be affected first: call centre staff, insurance underwriters, tax advisers, accountants, and legal secretaries are 98 to 99 percent likely to be computerised, according to Oxford University. In fact, PwC estimates that as many as 4 out of every 10 jobs in the US are at risk of automation.

In Africa, the picture looks slightly different. The continent’s large agricultural sector – estimated to account for 65 percent of all employment – relies on a vast network of smallholder farmers who till their own lands and sell their produce at local or regional markets. Until recently this was driven mostly by low-cost manual labour, but farmers are increasingly using digital tools to increase production and unlock opportunities to sell produce at large, often global markets. Digital transformation will not pass this sector by, and by effect AI will inevitably have an impact on even the most remote rural parts of the continent.

For larger enterprises operating in Africa, integrating AI into business processes holds the promise of yielding tremendous efficiency gains, while driving innovation and enabling new business models. It is not far-fetched to predict the rise of the intelligent enterprise – one where exponential technologies such as machine learning, IoT, blockchain and predictive analytics integrate into standard business processes to create entirely new business models – within the next 10 years.

However, the adoption of any exponential technology should not be driven purely with a profit motive. Technology must first and foremost support and enable human development and prosperity. And with an estimated 122 million people expected to enter the African job market by 2020 – coupled with the continent’s prevailing science, technology, engineering and maths (Stem) skills deficit – the rise of AI based automation should sound the alarm for businesses and governments across the continent.

The most likely immediate impact of AI is in its potential for transforming jobs and augmenting tasks. By liberating people from mundane repetitive tasks, they can focus on higher-value commercial activities that depend on abilities that are still unique to humans. Machines are not able to set goals or be responsible, and despite AI’s ability to extract value from data, that value is largely in past actions. Machines cannot match humankind’s ability to imagine, nor can they apply creativity, curiosity or emotional intelligence to any situation.

In Africa, the success of such exercises will depend on local access to digital skills. A reform of the education system is urgently needed, but may take several years – if not decades – to implement. The responsibility falls on all stakeholders to encourage and enable skills development in the Stem fields.

Governments and businesses also need to come together to develop guidelines for ethical AI to avoid the disenfranchisement of citizens and workers. In this regard, policymakers could utilise recommendations recently made to the European Commission.

Paige dos Santo is a digital transformation lead at SAP Africa.