Scientists don’t use AI, they use machine learning techniques to see if it can do something they were doing by hand more efficiently. If a scientist says they used AI, they are either lying or attempting to dishonestly win grant money, and neither is acceptable. Something called AI has no place in science. Not saying don’t have thoughts about this, but a pop sci article isn’t going to have sufficiently rigorous information to help anyone form a sufficiently informed opinion about pretty much any actual piece of scientific literature. I have also seen scientists waste many hundreds more good compute hours doing bad statistics with too large a data set or with a python script that breaks at the end of running to truly bad mouth attempts at using ML methods. It’s not particularly more irresponsible in and of itself than thoughtless and wasteful high throughput computing already is.
Scientists don’t use AI, they use machine learning techniques to see if it can do something they were doing by hand more efficiently. If a scientist says they used AI, they are either lying or attempting to dishonestly win grant money, and neither is acceptable. Something called AI has no place in science. Not saying don’t have thoughts about this, but a pop sci article isn’t going to have sufficiently rigorous information to help anyone form a sufficiently informed opinion about pretty much any actual piece of scientific literature. I have also seen scientists waste many hundreds more good compute hours doing bad statistics with too large a data set or with a python script that breaks at the end of running to truly bad mouth attempts at using ML methods. It’s not particularly more irresponsible in and of itself than thoughtless and wasteful high throughput computing already is.