I’ve often wondered whether insight is simply a particular kind of logic. Did Steve Jobs visualise the visionary Walkman by joining the dots and seeing an area where plenty of human activity took place yet recorded sound was missing? Was the iPhone an obvious step to take five years after Blackberry had proven the appetite for smartphones?

So it’s fascinating to contemplate the next big move that might, or will, come from artificial intelligence, after the recent launch of ChatGPT has revealed what can be done by analysing and regurgitation existing knowledge.

Make no mistake, ChatGPT is indeed a wonderful and wonderous technology and will have huge ramifications in the future. This blog should serve to demonstrate some of those implications because it couldn’t have been written by the technology.

Prompt: Can ChatGPT give insights about things that haven’t happened yet?

ChatGPT response:

“No, ChatGPT is not designed to provide insights about things that haven’t happened yet.”

Well, of course not, it’s a machine trained on countless millions of documents it found on the Internet. If it hasn’t been written about then it doesn’t know of it.

So it’s much more productive to ask ChatGPT about insight that already exists.

Prompt: List some of the insights people have had about the future of ChatGPT

ChatGPT response:

    1. ChatGPT will allow users to interact with AI-driven virtual assistants in natural language.
    2. ChatGPT will be capable of understanding the context of conversations and respond accordingly.
    3. ChatGPT will be able to generate natural-sounding responses that are tailored to the user’s interests and preferences.
    4. ChatGPT will use predictive analytics to anticipate user needs and provide personalized advice and recommendations.
    5. ChatGPT will be able to provide users with deep and meaningful conversations.
    6. ChatGPT will be able to learn from interactions with users and improve its understanding of language over time.
    7. ChatGPT will be able to provide automated customer support that is both accurate and efficient.
    8. ChatGPT will allow users to have a more natural, human-like conversation with a virtual assistant.

Generative AI regurgitates things that are already known. Its compelling feature is more a wonderment at the human-like language interaction, than it is about the existing knowledge it spouts. Indeed, it’s already been flagged as being a potential source of misinformation because it gets things wrong. So we must remember that while it’s a fascinating interface to existing knowledge, it has a way to go before it does a really good job at delivering that limited ambition.

Of course there is much hand-wringing beginning over the amount of human effort that such systems could do away with. The implications are little different from those confronting craftsmen in the early 1800s when the emergence of machines began to revolutionise the cloth industry, spawning the Luddite movement.

So might Apple’s new products team benefit from using ChatGPT in their quest for the next innovation?

Prompt: List some things that people want to do on the move that the iPhone can’t do

ChatGPT response:

  1. Make phone calls without a cellular connection
  2. Record high-quality video
  3. Play console-quality video games
  4. Send and receive text messages without a cellular connection
  5. Print documents
  6. Use a full-size USB connection
  7. Use a built-in GPS system
  8. Stream live video
  9. Use a physical keyboard
  10. Connect to a network via Wi-Fi or Bluetooth

This is hardly insight, it’s just regurgitated knowledge, some of it wrong, even if some of it is amusing because it’s obvious.

No, Henry Ford said: “If I’d asked people what they wanted, they’d have said ‘A faster horse’”, and he was right: existing knowledge has limited value when it comes to thinking of new ideas.

ChatGPT and other similar technologies are exciting and will bring about lots of change and new ways of doing things. The value of insight will go up as the skills required become less obscured by the drudgery of processing what we already know, but for now, the new ideas will have to come from humans.

Peter is chairman of Flexiion and has a number of other business interests. (c) 2022, Peter Osborn