I look at AI usage at work as basically taking on a bad but salvageable employee. For every use case, it needs a manager overseeing all their work and adapting to their strengths and weaknesses while also considering cost. It’s a deployment problem created by over promising.
We gave it a simple task of scan the internet for industry news, put results in a table formatted like x. It goes around 3-4 days before messing up. We have concluded that if it was an actual employee it would have either been sacked or put on performance review.
Yes, when employed properly it can be a helpful tool.
But when it’s given all the house keys and unlimited leeway, it will burn down your house (and your budget) because it cannot make reasonable choices. It’s not sentient (yet), despite all the promises from AI evangelists.
I look at AI usage at work as basically taking on a bad but salvageable employee. For every use case, it needs a manager overseeing all their work and adapting to their strengths and weaknesses while also considering cost. It’s a deployment problem created by over promising.
The salvageable employee will actually learn.
The AI doesn’t learn.
They’re only equivalent for the kind of manager who thinks investing in people is a waste of money.
We gave it a simple task of scan the internet for industry news, put results in a table formatted like x. It goes around 3-4 days before messing up. We have concluded that if it was an actual employee it would have either been sacked or put on performance review.
Yes, when employed properly it can be a helpful tool. But when it’s given all the house keys and unlimited leeway, it will burn down your house (and your budget) because it cannot make reasonable choices. It’s not sentient (yet), despite all the promises from AI evangelists.
It’s quite far from sentience, let alone sapience.