Sorry! I cannot help you with that

My last article on LinkedIn was over a year ago. It was about the rise of Generative AI. A lot of things changed in one year. AI models have become much more capable, and fast. And, the hype took a feverish turn, and now it is subsiding for good.

AI is extremely successful in the consumer space. Friends of mine who are not in the technology industry are putting AI to good use - to write, create, record, summarise. They can live with the inaccuracies and ambiguities of the AI responses. They themselves are human in the loop.

During the past one year, I have been testing business use cases of Generative AI. To put it simply, it works well as a co-pilot, not so much as an auto-pilot. It works well as an assistant, not so much as an agent. This is one reason the Generative AI is marketed through productivity and office suite, since there are often humans in the loop.

It may not be ambitious enough to use AI as a productivity tool. We need automated processes and autonomous agents, which deal with humans on the other side - colleagues and customers. The AI vendors realise the challenges well. So we have architectures and add-ons - to make AI work safely, efficiently and economically as much as possible. But would that be good enough? Is it okay if we are right 99% of the time? 99.99% of the time?

One approach that I took is to tone down the aspirations, by limiting the scope and capabilities of Generative AI. It is a new technology, a new paradigm, a new experience. We prioritise safety and reliability over surprise and excitement. The conversation with AI may not be exciting in the initial days. It is not superior to a conversation with a human.

And when it is not sure it says “sorry, I cannot help you with that”. And that is okay.