I agree with all of the above, so just to add some more info:
Basically the difference is that machine learning is a process and generative ai is a type of program resulting from such a process.
Machine learning is when a system learns "automatically", that is through some algorithm that a programmer has given it. That algorithm can be very simple or complex, for drawing a line between two points, or solving a sudoku, or for self-driving cars or writing a whole ass book, it's all machine learning.
AI is a program that has previously learned from or is continuously learning from machine learning. Most of the AI you see now is some combination of both.
Generative AI are such programs. They are programs that take an input and from that try to generate completely new unseen output. If you train it with 10 pictures of dogs and ask it for a dog, it will not show you one of the dogs or some weird blend of the dogs, but it will try to generate a completely new 11th dog using what it's learned about what a dog looks like from the 10 pictures. This also means that if the 10 pictures only included golden retrievers, it will never be able to generate a rottweiler. This is what data bias is. The person providing the data doesn't have to be aware of this bias. Maybe they were very cautious to include different hair types, fur colours, ear shapes or whatever and thought they did a really good job at avoiding bias. Point being that you can't avoid bias, but that doesn't mean that it's malicious.
Anyway, some examples of generative AI are the well known giants like ChatGPT for text and DALL-E for images, but it is also used for less flashy but nonetheless important purposes like Alphafold for protein folding!!! Figuring out the specific ways that proteins fold was a HUGE hurdle in biology for decades, and solving this problem helps solve way more problems than I can mention. I say this because I want to stress that "generative AI" is not the enemy. We're doing amazing things with it, it's just that the bad things tend to get more attention.
As for the other part of the post in the screenshot, that only generative ai is the problem? This is definitely very much absolutely not true. AI content moderation? not generative. Facial recognition? not generative. Targeted advertising? not generative.
There are so so so many different kinds of AI and there are so so so many different ways to use it. It is unfair to lump all of generative AI, or any other kind of AI, into the "bad" category. Doing so is dishonest and blinds us to identifying the real problems within AI, such as indeed biases, privacy, and accountability to name a few.
