As I noted in my previous article, artificial intelligence (AI) is driving many questions from business leaders as they seek to explore the full impact it can have on their companies.
A lot of concern has also been expressed about AI disruption. My advice on this point would be to prepare for disruption – plan for it. Before embarking on an AI implementation journey, assess the organisation’s readiness for it.
PwC advises that cutting across all these considerations is how to build AI in a responsible and transparent way in order to maintain the confidence of customers and wider stakeholders. But how to achieve this?
Research indicates that few businesses are ready to overcome the essential tasks necessary if they are to prepare correctly for such a radical change. For example, only 38% reported they can currently provide employees with reskilling and training opportunities in the face of technology disruptions.
Obviously, this is a serious misconnect and needs to be addressed. With the right strategy in place, it is possible to prepare the organisation and its staff to not only survive automation but thrive on it.
AI technologies permit businesses to mine data, generate insights, create operational efficiencies, provide stronger experiences, and close the gap between information and action in ways previously not possible.
All companies want to avail of this potential power, which also unleashes the ability to disrupt, innovate, enhance, and in many cases, totally transform an organisation. AI is an umbrella term encapsulating machine and deep learning, image and video recognition, predictive analytics, process automation, speech recognition, biometrics and natural language processing. It can apply to practically every industry sector.
What is AI’s value proposition in 2022?
Forrester’s predictions for the use of AI in 2021 and going forward are clear on how AI will influence business development in 2022.
The research company says AI and machine learning will permeate new use cases with companies pushing it to new frontiers, such as holographic meetings for remote work and on-demand, personalised manufacturing.
It goes on to highlight how AI is expected to boost workplace automation and augmentation needs.
It is possible to prepare the organisation and its staff to not only survive automation but thrive on it.
In 2021, a third of companies in adaptive and growth mode looked to AI to help with workplace disruption for location-based, physical, or human-touch workers, and knowledge workers operating from home. These businesses were reported to be applying AI to intelligent document extraction, customer service agent augmentation, return-to-work health tracking, or semi-autonomous robots for social distancing.
Gartner predicts a robust AI engineering strategy will facilitate the performance, scalability, interpretability and reliability of AI models, while delivering the full value of AI investments. It notes AI projects often face issues with maintainability, scalability and governance, which makes them a challenge.
It states that AI engineering offers a pathway that makes it part of the mainstream DevOps process rather than a set of specialised, isolated projects. It brings together various disciplines to tame the AI hype, while providing a clearer path to value.
Gartner highlights the fact that due to the governance aspect of AI engineering; responsible AI is emerging to deal with trust, transparency, ethics, fairness, interpretability and compliance issues – in other words, AI accountability. This drive towards trusted data and how it is used is particularly pertinent in SA.
AI and POPIA
No discussion on AI in South Africa would be complete without examining POPIA – which came into effect on 1 July 2021 − and its impact on the use of AI systems.
South African businesses will be remiss if they do not familiarise themselves with a deep understanding of the regulations pertaining to the use of personal information required for AI or acquired through it.
The Act has made provision for the governance of automated decision-making and strictly directs how pronouncements or resolutions may be arrived at through information gleaned from AI profiling techniques. It prohibits financial institutions, for example, from granting or rejecting loan applications solely based on profiles created by AI systems.
Ignorance will not be accepted as an excuse if businesses inadvertently put themselves at risk of breaking with regulations. Analytics lies at the heart of AI systems that produce information which may be deemed valuable to the system but may not have been compliantly acquired as directed by POPIA.
Add to that the risk of a security breach which today is being cited more as a certainty than a possibility. In a case where personal information is compromised and data is stolen, the organisation that gathered it will be left with a fine and possible staggering loss of business due to breach of trust with its customers.
Do these constrictions mean AI cannot be used to tap into the unparalleled value it holds for businesses today? No, it certainly does not, but it does require partnering with AI specialists who can guide companies through the regulatory minefield to ensure compliance.
In my final article in this series, I will discuss the biggest fear with AI: the possibility of job losses due to automation and AI.