I agree 100%.
What I want most to study is AI and Data Science. However, I’m okay to study CS / Engineering as an undergrad and then study AI / DS in grad school. I’m not 100% sure about economics but I still applied in case I change my mind. (those programs were numerus fixus and they don’t accept applications after Feb).
What do you mean by “CS / Engineering”? Do you mean Computer Engineering (which is different from Computer Science)? AI is a subfield within CS. Data Science may or may not be a part of CS, depending on the school. It overlaps with CS in a number of areas, including its use of AI machine learning techniques.
I still believe the LACs offer the best flexibility and preparation for graduate school.
I normally would agree. However, in this case, admissions to top graduate programs in AI are even more competitive than getting into a tippy-top undegraduate college. For example, acceptance rates to a few top AI PhD programs are at much less than 1%. Significant experiences in research, and even publications, are necessary, which make LACs less than ideal.
LACs are actually very good for undergraduate research and publications, since one criterion for being hired as faculty is strong research and ability/willingness to work with undergraduates on that research. (There are no graduate students who’d have priority).
However OP needs to check what kind of research the CS faculty is doing at Denison. In addition, a very strong background in Math/Statistics would be expected, so that the ability to add a minor in these subjects (or equivalent) should be taken into account.
The types of researches carried out at different types of schools are often very different. My son, a college junior, has been doing a year-long research in the same subject area. The professor told him about two years ago that he would be given this opportunity only after he’s taken certain background graduate-level courses. For his PhD application, he must also have more or less completed his research by the end of his junior year. LACs are great options in some cases, but not in this one. They simply offer neither the right types of courses, nor the right types of researches.