As an AI researcher, I think a lot of people are a little too sensitive to the term "AI" and make a lot of big assumptions upon hearing it. It's a very general term that doesn't really imply any particular degree of complexity or sophistication. Labeling simple machine learning algorithms and heuristics as "AI" isn't at all unique to this era of hype that began in the last ~5 years -- rather that's how the term has been used in academia for many decades. If you took a college class called "AI" or looked up some of the most popular textbooks on AI [1], you'd find that a lot of it is dedicated to search algorithms (breadth-first, depth-first, A*), linear classifiers, and feature engineering. If you think "artificial intelligence" is a bad name for these things, fine -- but don't blame the recent wave of hype, this is what the term AI means and has pretty much always meant. So go ahead and call your startup's linear regression "AI", and if the VCs leap to fund you under the impression that it means you'll be behind the singularity, that's on them. AI != deep learning. AI != AGI.
[1] e.g., "Artificial Intelligence: A Modern Approach" by Russell and Norvig