This may be a stupid question but I couldn't really find an answer on the internet. I am a 3rd year math major who is interested in computer science, particularly algorithms and competitive programming (did some olympiads in high school, ACM ICPC in university, etc.) and I have been meaning to get into AI. I have all the prerequisites to get started but the problem is that I really, really hate statistics. I took a course on it last year and found it to be very dry. I've heard people say that AI is mostly statistics and I am very concerned if it's true. I can tolerate some amount of stats but if the field literally revolves around it I will not be able to do it. So, exactly how much statistics is involved in AI? Are there fields of AI which use it less than others?
Many people without a formal/solid background in statistics (e.g. without knowing exactly what the central limit theorem (CLT) states) are doing research on machine learning, which is a very big and fundamental subfield of AI that has a big overlap with statistics, or using machine learning to solve problems.
So, in my view, you don't need to learn everything about statistics to do research on some AI topic, including machine learning, but you need to have an understanding of the basics (at least a full introductory college-level course on statistics and probability theory), and the more you know the better.
More specifically, if you don't know what the CLT or the law of large numbers state, you will not have a full understanding of many things that are going on. At the same time, you will find a lot of research papers (published in ML conferences and journals) that do not even mention hypothesis testing, but it's important to have an idea of what a sample, sample mean, sample variance, likelihood, maximum likelihood estimation (MLE) or Bayes' theorem are. In fact, MLE is widely used in machine learning, but not many people using/doing ML would probably be able to explain precisely what the likelihood function is.
Finally, in my opinion, having a formal/solid (not necessarily extensive) background in statistics should be a prerequisite for doing research in machine learning (you need to really know what the likelihood function is!), which some people called applied/computational statistics or glorified statistics for some reason, but not necessarily for using machine learning to solve some problem. Moreover, there are other areas of AI that do not make use of statistics, but ML is probably the most important area of AI. So, if you hate statistics, you may not like AI and particularly ML, but maybe you will change your opinion about statistics, once you understand what e.g. neural networks are capable of doing or not.
I work in NLP, and use very little statistics. Actually, almost nothing I do can be classed as 'serious' statistics.
So yes, AI is a wide area, and in my company there is a group that does machine learning, so they probably use a lot more of it than I do. Previously I worked in conversational AI. Again, very little to no statistics at all.
I would contest the view that AI is intrinsically data-driven. That's one aspect of it. However, while I look at actual data (texts) to derive algorithms for their analysis, I don't need to use any statistical concepts for that. And even evaluation of the results is just counting and comparing.
There are statistical algorithms in NLP, but they are not usually very complex or hard to understand even without a lot of stats knowledge.
Thanks for the extra details. There area good answers already, but I'll give just a bit more information since your requirements are a little more specific now.
Since you mentioned Research Engineer only, I'm going to assume you are not really interested in a plain engineering role.
I can say for a specific Research Engineer role I am aware of at a world class industrial AI lab, their minimum requirements include "calculus, linear algebra, and statistics at least to a first year degree level". So it sounds like you already have this required level if you were to apply today.
On the other hand, I would be cautious regarding what you found dry about your stats course. If thoughts such as "this portion of the dataset have somewhat less representation in the results and this other portion have somewhat more, I wonder why that is?" sound very dull to you, you may not like it. Most current AI is based on large sets of data. I am referring to deep learning / neural networks here rather than previous methods, but that is where a lot of the hype / major breakthroughs are at the moment. In computer vision which you mentioned for example, current methods typically input a large dataset of images to create the AI system, then test it on a large dataset of images. If they are images of road signs, you might find that increasing the proportion of one type of road sign makes the system worse for another type of road sign. Identifying relationships like that is an important part of the research function. The more towards the research side rather than the engineering side you are, the more you will need to be able to analyse things like this yourself.
Despite all that, I found stats to be one of the drier maths subjects, yet I very much like AI (and stats in AI).
“When you're fundraising, it's AI. Statistics is the field of mathematics which deals with the understanding and interpretation of data. ... Machine learning is nothing more than a class of computational algorithms (hence its emergence from computer science). Artificial intelligence (AI) is intrinsically data-driven. It calls for the application of statistical concepts through human-machine collaboration during the generation of data, the development of algorithms, and the evaluation of results