I think alot of the replies here have selection bias from people who have had a pretty smooth road, or already have the maturity to handle it.
For many, a PhD is an arduous journey with a steep learning curve and maturation journey; alot of bad habits have to be reliquished; alot of noise has to be filtered; a purity of purpose has to be learned.
Moreover, doing a PhD "for the purpose of being one of the superstars who make a breakthrough" is probably a really unhealthy goal because of the 20000 papers submitted to each of the large conferences each year, a tiny tiny fraction can be called that, if that. Research is inherently open ended and by its very nature, past success is not necessarily predictive of future success, especially in the "breakthrough" category.
During a PhD one often watches friends of the same age get a faster start in life, accumulate wealth at a faster pace, etc etc which in maslow's hierarchy of needs, means putting alot of human needs on hold, resulting in the tremendously high depression rates among students.
The ONLY reason, the sole reason, I think one should contemplate seeking a PhD, is love of the nature and natural curiosity above all else 🌹.
I think that all other reasons that one might think to go in with (fame? immortality? wealth? stardom?) will rapidly fade as the difficulty of feeling Nature's pulse sets in. All these other reasons probably have quicker and better paths to success, anyway, like going on tiktok or elevator pitching a VC.
Should you get a PhD in CS/AI?
Both seem to be true:
1. 95% of PhDs would have done equally well career-wise without one.
2. 95% of real AI breakthroughs (ImageNet, Transformers) came from PhDs.