I edited your question to make the language more precise because they're just computer programs. Here's my question and the answer you wanted:

Q: What is the mathematical average of the ages of rock stars performing concerts?

AI Overview
Based on concert revenue reports from 2019, the average age of rock stars and other top-grossing musicians on tour is around 53 and rising. This number is heavily influenced by the continued touring success of "legacy" artists who have been performing for decades.

Factors influencing the high average age

Established success: Older, classic rock acts like the Rolling Stones, U2, and Metallica have a massive and dedicated fanbase that can afford higher ticket prices, ensuring their tours are consistently profitable.
Older fans: The fanbases for these legacy acts are also aging and tend to have more disposable income to spend on high-priced concert tickets.

Statistical impact: The presence of a few very old, highly successful touring artists can significantly skew the average age upward, overshadowing younger rock acts. For example, in 2019, artists like Elton John (72) and Fleetwood Mac (with members in their 70s) were among the top 10 touring acts.

20 + 20 + 50 + 70 = 160, not 170.

Further, the answer given does not contain 20, 50 or 70.

and changed one of the 20s to 30, so 170.

The ai response I noted mentioned 20s and 30s, some rockers in their 50s and some over 70. I used those number to make an example of how to average. I'll be sure to go get an eye test, and then return to 4th grade math.

So the AI mentioned some decades. Decades are much broader than ages.

It did not provide anything for you to assume those decades should be equal-weighted. When it mentioned those decades it mentioned them. It did not spew out vast quantities (one key connotation of "spew" ) or blurt them randomly in a spotty uncoordinated way (another key connotation of "spew" ). I think that you are approaching this with confirmation bias, as indicated by your use of "spew" in a pejorative way.

There may be many people, different numbers of people in those decades. Simply adding up the decades with no account for the number of their members is a very inadequate approximation or model of the reality.

Your result of 42.5 is not an indication of stupidity on the part of the AI. It is an indication of the inadequacy of your model. This is proven by the evidence quickly provided by other posters here.

If you want to make a case for "AI" being stupid, you have to dig up actual data, actual information and then process it in the way you ask an AI to process something. Rolling Stone magazine did the math.

Searching Google for "average age of rock stars performing concerts" (without the quotes) yielded this AI Overview:

Am surprised that it didn't take you literally.