Yes excellent essay. Basically one of the core values of progress of any kind is to detect error. "Ignorance is not the enemy of progress. It is the illusion of knowledge." Daniel J Boorstin. See also Hawking.
The most influential paper in the AI world, ironically, is titled “Attention Is All You Need‘’ (2017 original/2023version7). All computer scientists of note, should be aware of this particular paper, and naturally, of many others as well. AI is a highly intimidating environment, as evident (for example) from the vast domain of journal papers at the open-access archive for scholarly articles, arXiv as managed by the USA-based Cornell University. I intentionally refrain from providing my personal views on this policy, as it has so many shortcomings, and is in a most serious lack of substance and relevance, and inaccurate in so many ways, and ‘’senseless’’ in many of its assumptions, and having no regard to the current risk-matrix-profile so clearly voiced by notable spheres, that one just want to conclude by saying: the policy in its current frame-of-mind is an utter disgrace in terms of academic excellence, and with due respect, will not even pass a Senior Certificate-examination. And – of course – apartheid is blamed again in this policy if AI will widen the gap of inequality in South Africa. The ‘mirror’ to transform mindsets has again been placed towards the past, and not to leadership themselves, which is their modus-operandi for many years now.
As regards the AI Policy, Johan, a more appropriate title for your post might be: 'I cheat, therefore I am.' You seem reluctant to use the word 'cheat', but one obvious thing about AI is that it makes cheating much easier. And it provides a ready-made excuse: 'The LLM hallucinated.' In other words: 'The machine made me do it.'
Johan, as usual I am greatful that you took such great interest in this "human" debacle. The danger of not putting it into perspective, is that AI per se will be the easy "accused" to put on trial. That is the lazy way out, but then - just like your example of the students' "human error excuse" - most other people (and not only students!) often choose the lazy view out. Your sensible writings will provide the necessary balanced context, as you have already started doing in the article, and will continue doing in further upcoming material on the subject, as you intimated. We MUST, especially in the South African context, guard against setbacks to large-scale AI application on account of naive utilisation as per the IT Policy "oopsie". That was surely human error/-misjudgement/-inexperience which was at the foundation of what happened. BUT HERE WE MUST TRY HARD TO PREVENT THE LAZY "SOLUTION" TO NOW THROW OUT THE AI-BABY WITH THE BATHWATER. If that happens in especially bureaucratic circles, on accout of "fear of AI mistakes", we run the risk of putting ourselves as a country (again) many years behind other countries...
Yes excellent essay. Basically one of the core values of progress of any kind is to detect error. "Ignorance is not the enemy of progress. It is the illusion of knowledge." Daniel J Boorstin. See also Hawking.
The most influential paper in the AI world, ironically, is titled “Attention Is All You Need‘’ (2017 original/2023version7). All computer scientists of note, should be aware of this particular paper, and naturally, of many others as well. AI is a highly intimidating environment, as evident (for example) from the vast domain of journal papers at the open-access archive for scholarly articles, arXiv as managed by the USA-based Cornell University. I intentionally refrain from providing my personal views on this policy, as it has so many shortcomings, and is in a most serious lack of substance and relevance, and inaccurate in so many ways, and ‘’senseless’’ in many of its assumptions, and having no regard to the current risk-matrix-profile so clearly voiced by notable spheres, that one just want to conclude by saying: the policy in its current frame-of-mind is an utter disgrace in terms of academic excellence, and with due respect, will not even pass a Senior Certificate-examination. And – of course – apartheid is blamed again in this policy if AI will widen the gap of inequality in South Africa. The ‘mirror’ to transform mindsets has again been placed towards the past, and not to leadership themselves, which is their modus-operandi for many years now.
As regards the AI Policy, Johan, a more appropriate title for your post might be: 'I cheat, therefore I am.' You seem reluctant to use the word 'cheat', but one obvious thing about AI is that it makes cheating much easier. And it provides a ready-made excuse: 'The LLM hallucinated.' In other words: 'The machine made me do it.'
Johan, as usual I am greatful that you took such great interest in this "human" debacle. The danger of not putting it into perspective, is that AI per se will be the easy "accused" to put on trial. That is the lazy way out, but then - just like your example of the students' "human error excuse" - most other people (and not only students!) often choose the lazy view out. Your sensible writings will provide the necessary balanced context, as you have already started doing in the article, and will continue doing in further upcoming material on the subject, as you intimated. We MUST, especially in the South African context, guard against setbacks to large-scale AI application on account of naive utilisation as per the IT Policy "oopsie". That was surely human error/-misjudgement/-inexperience which was at the foundation of what happened. BUT HERE WE MUST TRY HARD TO PREVENT THE LAZY "SOLUTION" TO NOW THROW OUT THE AI-BABY WITH THE BATHWATER. If that happens in especially bureaucratic circles, on accout of "fear of AI mistakes", we run the risk of putting ourselves as a country (again) many years behind other countries...