The Guardian Weekly of 4 January 2019 carried an article by Vivienne Ming about Artificial Intelligence. In theory, she says, poverty, mental health, climate change, inequality – almost everything – could be addressed by AI. The problem, she says, is not the concept of AI, but the people behind it. She points out that AI is being developed by young men (mostly) “who have never solved a problem in their lives. They have never done anything from scratch to make someone’s life better”. And here we are trusting to these youngsters (some of whom are technically brilliant, but possibly on the autistic spectrum, (although she doesn’t mention this, nor do I have any data to prove it) to usher us into the Age of Artificial Intelligence.
Take the issue of gender bias. Some techies seem to think that they can throw an algorithm at a subject like this and it will come up with an answer. But if the people who constitute a company don’t know how to avoid bias in real life, AI will not solve the problem. If you throw a neural network at a pile of data, it will find patterns that predicts a person’s grades, job prospects, or the odds that they will re-offend. But human beings are infinitely complicated. What we desperately need is deeper understand about life – the real causes of good grades, re-offending, or why a person seems to have good job prospects.
The answer is to recruit advisors who are old, experienced and who have a measure of wisdom. It’s the old story – if you feed in trash out will come – trash.