In short, the broader the use case and more general the AI deployment, the riskier the outcomes are likely to be. Using a large language model to summarise or convert content is a good use case. Offering an AI-powered general purpose chatbot to your customers is likely to cause a plethora of problems, particularly where you are not able to vet the data that the model has been trained on. Businesses need to make sure that they aren’t getting themselves into sticky situations by moving too quickly or deploying technology that can break in unexpected ways. Governments and regulation moves slowly, but at some point it will catch up and there may be a lot of companies exposed to lawsuits and fines when it does.
This book is required reading for everyone involved in looking at Generative AI for their business or in their work.