This book by Professor Katherina A. Zweig provides an overview of AI, especially on the caveats of depending on its outputs and the need for ethics when drawing on AI as a collated / sifter / decision maker.
Of note is an explanation of the computer science behind the algorithms which underlie AI. This is of especial importance as new generations of AI, exampled by ChatGPT and the incorporation of this technology into search engines - see the article by the Guardian on Google and Microsoft's swift moves into integrating AI. The main caution is that the algorithms governing how AI generates it's outputs, are far from perfect. Decisions made on outputs from prompts entered into the system contain bias, as the outputs are only as good as the data from which the outputs generated are made.
The argument in the book is that AI may be best for situations where things are 'black and white' eg. as in playing chess or go. Given the inherent fuzziness of the human condition and the greyness / continuums of things for may situations, AI may not be able to respond authentically, ethically or creatively to challenges which do not have clear cut answers.
No comments:
Post a Comment