I like how Sandra Matz expresses the idea of life with AI.
“Statically Safe Sameness,” she says as she expresses the concept of safe exploitation, decision-making, and exploration.
AI always chooses the safe options, flattening and removing creativity.
Not that it is not capable, but because it lacks incentives.
AI is rewarded for choices that people like. So, if it suggests ideas that are creative and people are not rewarding it, AI will play it safe for reward.
She talks about her experimentation. Asking AI to choose an ice cream flavor from a long list of options. Her explanation here is truly eye-opening, worth watching.
TED Talk link: Is AI Turning Us All into the Same Person? | Sandra Matz | TED
She asks us to use and reward AI, to build with AI in a way that preserves human complexity and allows AI to be a bit more creative, building on our preferences. So not everyone becomes a “basic bitch” by following AI’s agenda.
AI is the new “Can’t you just Google it!”
I agree with her message. Everyone is using AI nowadays. When Google became mainstream, everyone started searching on the web and everyone famously (annoyingly) said: “Can’t you Google it?” It was almost like people were irritated to answer something that is common enough knowledge to “Google it”.
Now technology went a step further and made information not just searchable, but seemingly intelligent, interactive, and self-acting, if I dare to say.
Now everyone asks AI. Without validating what it says, most people accept its personal advice and knowledge.
Asking a general contractor what it would take to renovate my house and how I can finance it? He asks AI.
My wife is asking if the battery we have is for a gift tool she wants to purchase for me? Asks AI.
My friend asking AI how to get back data from his laptop that was splashed with fluid and stopped working? When I tell him from experience what it takes to do, he still prefers AI’s answer.
The result is on us
I don’t mind, but I know its answer has flaws. Not because it lacks information, but because what we ask is missing details, and AI won’t ask back for more information (at this time of writing), but will presume based on a general presumption or “Statically Safe Sameness” and will answer based on that.
By the way, we will call expert subcontractors to do the real structural assessment, and do what it takes to get a project that scale done.
About my wife’s advice. No, the battery I have is not working with the tool she bought me. Nor the one she was thinking about to buy 😀
And about my friend with his laptop. He still has to take it to a real laptop service and a data retrieval specialist; neither my advice nor AI will help get it done.
I might have drifted from the original message of the TED talk in terms of topic, but it is all connected, and these are the thoughts and experiences triggered by that video.
How about you? What are your opinions?
