When calculators were first introduced, there were several concerns, particularly in the educational sector. Many educators and parents feared that students would become too reliant on calculators and lose their ability to perform basic arithmetic operations manually, leading to a decline in mental math skills. There was also a worry that calculators would erode students’ understanding of mathematical concepts, allowing them to rely on technology without grasping the underlying principles. Equity in education was another issue as early calculators were expensive and not all students could afford them. Concerns also included potential cheating and resistance from educators who preferred traditional teaching methods over adopting new technology.
Despite these concerns, calculators eventually became widely accepted and are now considered an essential tool in education, helping to enhance learning and understanding of more complex mathematical concepts. This sounds exactly like the concerns we are raising for the introduction of AI. “But, this is different” I can hear people say. Yes, it is undeniably different and even more pervasive. Here are some examples.
Medical AI
In 2001, I was diagnosed with Chronic Myeloid Leukaemia and began treatment with a groundbreaking medication called Gleevec. Remarkably, all signs of cancer vanished. Accessing the medication had initially been a challenge because my doctor, a bone marrow transplant specialist, was not attuned to the latest advancements in targeted therapies. While this was a miraculous development, my doctor unfortunately passed away and all my medical records were lost. The new doctor needed to renew my prescription but found no evidence of the cancer. Thankfully, the medical aid had retained the necessary proof.
A robot equipped with AI would have retained all my records, tracked the progression of cancer over the years and been aware of the developments in CML treatment. It would not have pursued the marrow donor route, which turned out to be a failure since no match could be found.
ChatGPT
ChatGPT, developed by OpenAI, is a product of extensive research and development in the field of artificial intelligence, specifically in natural language processing. Its foundation is the GPT architecture, based on the Transformer model, which is a type of deep learning model designed to understand and generate human-like text. It uses self-attention to better understand context and relationships in language. Its development involves pre-training on large internet text to predict the next word, followed by fine-tuning with narrower datasets and human feedback to improve accuracy and alignment with human values. OpenAI has released several iterations of the model, improving size, capability and safety. Ethical considerations such as bias reduction and responsible use have guided development. ChatGPT is now widely accessible for tasks ranging from conversation to complex problem solving. Overall, its development is a testament to the advancements in AI research and the ongoing efforts to create models that can effectively and safely interact with humans.
Any person who has worked with a program like ChatGPT would know that one cannot accept everything that is produced at face value. It pulls information from the internet and has rules that prevent it from doing certain things. I asked ChatGPT who was right: Russia or Ukraine. I got a long discourse that boiled down to “it all depends”. To know who is right and who is wrong takes more information and logic than what ChatGPT can find on the internet. The answer one gets from ChatGPT depends on how one phrases the question. If you know how to ask and what to ask for, you will succeed. If you don’t, your competitor will succeed and all you will get is an answer saying “it all depends”.
Automation is here to stay and will take over more and more mundane as well as dangerous tasks. Video game players will replace soldiers and pilots, robots will take over transportation of goods and 3D printers will be used to manufacture parts. In South Africa, we are worried about automation taking over our jobs. This concern is valid as automation will definitely take over some jobs, but there are things that it cannot do – yet.
Skilled trades like electricians, plumbers and carpenters rely on manual dexterity and problem solving in unpredictable environments, making them less susceptible to automation. Management and leadership roles involve strategic decision making, people management and understanding organisational culture, which are difficult to automate. Social and community services, including social workers and counsellors, require empathy and insight into complex emotions. In education and training, teachers provide personalised learning and mentorship, which AI can support but not replace.
Since most of the readers in this magazine are technically minded, they are relatively safe. However, keep yourself updated with the latest technologies otherwise your replacement will. This is why CPD events are so important. We need to make sure we keep ourselves updated and prepared for the changes coming, changes that are so drastic that they will shake the foundations of those who are not prepared. Technology is moving fast and we need to keep up. The rules regarding professional registration are not to make our lives difficult, they are there to play a small role in assisting us to remain relevant.
Yours in automation.
| Tel: | +27 11 312 2445 |
| Email: | ina@saimc.co.za |
| www: | www.saimc.co.za |
| Articles: | More information and articles about SAIMC |
© Technews Publishing (Pty) Ltd | All Rights Reserved
printer friendly version