It really depends on what you mean by AI. If by AI you mean the standard LLMs everybody uses day to day then you’re probably right for a while.
But there’s more to AI than just the chat interfaces we’re used to. For example DeepMind published some research on how they trained a system that leveraged LLMs in a unique way and found it was capable of solving novel math and computer science problems [1].
DeepMind also trained AlphaGo which famously showed creativity in its game against Lee Se Dol that puzzled experts who initially thought it was a blunder.
Don’t forget that just 3 years ago everyone was debating whether a machine could pass the Turing test and today nobody even thinks about that anymore.