Much like "old" seems to be "your age + 15", then AI seems to be "whatever algorithms aren't yet well established". This relative view of AI is a misunderstanding with negative consequences that I'd like us to correct.
For example, this HBR article claims AI failed us in the pandemic. Why? Because some attempts to diagnose and predict COVID using image and sound analysis failed. Of course, a more accurate title would be: Technology worked impressively and there was also experimental research. The author ignores that, for example, the technology that allowed the world to know the COVID genome sequence in days(!) was developed in the same computer science, signal processing, and statistics, research departments where current "AI" is supposedly "failing us". The HBR article seems to also forget just how much the efficient deliveries that kept a lot of the economy moving, were developed in the initial years of AI research. As algorithms succeed and find applications they become normal, fade into the background and get demoted from being called "AI" - they become "technology".
A focus on whatever experimental application is grabbing the headlines, or whatever new commercial product is less than a few years old creates an easy (moving) target for criticism. Ignoring the mountain of accumulated success of the past decades (now out of the media hype) creates a public perception of technology that is ill-informed. This harms people: it creates misunderstanding and potential mistrust and misregulation of technology that is already saving lives.
I was once asked in an education conference "when will AI change education?". It's already changing education! Don't focus on technology names, think of applications: think of the hearing impaired student who can now read transcripts of the class; or think of the virtual classroom where noise cancelling systems that help everyone focus better. The technology that powers speech transcription and vocal isolation are the result of the AI that is "failing us". All increasingly at the reach of everyone with a smartphone or an internet connection.
The research edge of technology will always have lots of dead ends, and lots of successes. Focus on understanding particular uses cases, applications, and what safeguards we need - regardless of what people call it.