How to build trust in the age of AI

0
175

Ian Nyamusa | Senior Manager | KPMG South Africa | mail me |


 

 

 

 

 

 

 

 


Rovanya Kasaval | Analyst | KPMG South Africa | mail me |


Imagine cruising down the highway in your Tesla (Model 3), sold by the dreams of the autopilot feature. Suddenly a disaster strikes as the car’s AI fails to distinguish a semi-truck ahead, resulting in a fatal accident. This scenario that occurred in Florida on March 2019 highlights the potential dangers of reliance on AI in autonomous vehicles. Yet, in October 2023, researchers from the University Medical Center Utrecht, developed a deep learning algorithm; a form of AI, to identify brain tumours, speeding up diagnosis which now goes on potentially save millions of lives.

These contradicting realities – disaster and lifesaving – depict our complex dance with AI. This article explores the ethical considerations, pitfalls, and safeguards needed to ensure AI becomes our guardian angel rather than humanity’s downfall.

While AI might sound like science fiction, its roots stretch back to the 1950s at the Dartmouth Conference in Dartmouth College in New Hampshire, USA, where the term AI had yet to be known. Following this conference, in 1956 Arthur Samuel had a vision to implement the first alpha-beta pruning algorithm in an IBM 704 computer that would determine the probability of winning a chess game.

Over time, different subsets emerged:

  • Machine Learning (ML) – (1997) for data-driven pattern recognition,
  • Natural Language Processing (NLP) – for understanding human language, and
  • Deep Learning – (2017) mimicking the brain for complex tasks.

These specialised forms led to Narrow AI excelling in specific areas like facial recognition. Now, the horizon holds the potential of Generative AI (Gen AI, ChatGPT, November 2022), capable of producing content (e.g. text, images, videos) in response to prompts with human-level intelligence across domains.

Trust

But can AI be trusted?

Although the horror stories of robots gone rogue might capture our imagination, South Africa is currently amidst the positive impact of AI implementation.

South Africa’s Big 4 Banks are creating strategies to facilitate the adoption of AI in the financial sector. First National Bank has embarked on a journey of AI, ML, and Gen AI to combat fraud detection and improve risk management. While ShopriteX; a business unit dealing with Emerging Technology by Shoprite has started leveraging AI and ML tools to address food wastages and logistics issues. Discovery Health employs sophisticated AI on their Dr Connect App providing reliable answers to members with medical symptoms.

It is without doubt that AI can create a better, automated future but the ethical considerations surrounding its development and use, require careful principles and practices. Can we truly trust AI? – the question we must answer as we move forward in this age of artificial intelligence.

The undeniable benefits of AI initiate a crucial understanding in acknowledging the potential risks associated with its implementation. Every organisation participating in the realm of AI should prioritise safeguarding measures to mitigate risks.

Five crucial risk areas to consider are elaborated on below:…


The full article is reserved for our subscribers!

Read the full article by Ian Nyamusa, Senior Manager and Rovanya Kasaval, Analyst,  KPMG South Africa, as well as a host of other topical management articles written by professionals, consultants and academics in the April/May 2024 edition of BusinessBrief.


VIEW our subscription options

ALREADY SUBSCRIBED?


Questions or problems?

admin@bbrief.co.za | +27 (0)11 788 0880 |


 



LEAVE A REPLY

Please enter your comment!
Please enter your name here