Linked – Have the Right Expectations Regarding Artificial Intelligence
I want you to pay particular attention to this example of using AI to come up with better icebreakers for meetings:
But neither of them created anything new. Both icebreaker activities have been around for decades. They are not original. I’m not saying either one of them is bad. I am saying that I was surprised that AI suggested something so … well, old.
Sharlyn chalks this up to AI still learning, but I don’t believe that’s correct. I don’t think Generative AI tools can come up with anything new. That’s not the way it works. Large language models are developed by ingesting and analyzing tons of existing data. Based on that data, they then mathematically conjure up the most likely response to your prompt. How could it come up with a genuinely original icebreaker?
Unless I’m misreading how LLMs work, I don’t see how AI could respond to any prompt with something that hasn’t been done or written down. Expecting that AI will create something new is what we need to change. At most, and it’s not very good yet, it can repurpose some ideas into something that looks new but is derivative of something that existed before.
Of course, if you believe that there is already so much creative content in the world that everything is derivative now, then maybe it doesn’t matter. Perhaps that level of AI is doing precisely what we are doing. We like deluding ourselves into thinking we are creating original ideas when they are all derivative.
In the meantime, though, accept that AI is not likely to present you with something no one has thought of before, and try to use it to find things that you’ve not thought of before.
Follow these topics: Tech
