LETTER: Elephant in AI room

Johan Steyn wrote an illuminating piece on the need to balance “ethical stewardship” and “innovation” in the development of AI (“Beyond the code: the ethical dilemmas of uncontrollable AI”, September 18). It is a praiseworthy objective. However, this is already being done by the widely used Meta AI programme.

A current ethical issue is whether it is ethically desirable to slaughter 200 elephants to feed hungry people, as proposed in Zimbabwe. To see if AI can do what Steyn suggests, I asked Meta AI the question. It gave a swift reply.

The answer was that it is not ethically desirable to slaughter 200 elephants for the benefit of hungry people. While providing food is a moral imperative, the long-term consequences of harming elephants can destroy conservation and ecological benefits, which are for the long-term benefit of many others. These outweigh the short-term benefits, such as sustainable food.

However, it added that AI has limitations. It lacks human empathy and emotional understanding and has potential bias in its ethical framework. AI provides valuable insights but should not replace human decision-making in complex ethical dilemmas.

AI does not answer ethical issues. It provides a framework for doing so. I suggest Business Day readers test this by asking AI some thorny ethical problems.

Andrew Prior
Newlands

JOIN THE DISCUSSION: Send us an email with your comments to letters@businesslive.co.za. Letters of more than 300 words will be edited for length. Anonymous correspondence will not be published. Writers should include a daytime telephone number.