Intelligence: Artificial vs Biological
While everyone knows roughly what is meant by intelligence, it is hard to pin down a general definition. Part of the reason for this is that it is used differently in different communities. Attributions of intelligence to biological creatures do not match those to artificial creations. One possible explanation is that artificial intelligence tries to solve a specific problem (or set of problems), usually starting from scratch each time, which causes intelligence to be associated with learning (often from scratch). On the other hand, biological intelligence takes already existing entities and looks at how well they adapt to uncommon (for them) tasks.
To explore these claims, and the differences between the two ways of understanding intelligence, I scraped papers mentioning intelligence (and variations on the word) from both arXiv (an archive for papers in physics, mathematics, computer science and related areas) and biorXiv (for papers in biology). While it's hard to draw any real conclusions from such an approach, it does seem to support the main difference being the focus on learning.
Intelligence and Learning
There are many issues with this type of informal analysis, and any conclusion should be taken with a grain of salt (if taken at all). However, the prevalence of learning in the arxiv data is quite striking. 'Learning' doesn't show up until the 70th most common in bioRxiv. One possible explanation is simply that deep learning and machine learning are common terms (both 'deep' and 'machine' also appearing highly in the data). However, removing all instances of 'deep learning' and 'machine learning' from the data still leaves 'learning' as the third most common word on arxiv (after model(s) and data).
Perhaps there is a real difference here, that will interfere with any attempt at a general definition of intelligence applicable to both biological and artificial systems. Perhaps it is a sign that AI is still in its infancy, and, as it progresses, will move away from a focus on learning from scratch, to a focus on improving performance across tasks on systems already built or trained to be competent at others. Perhaps it is the other way round, and the preceding leans too hard on biological systems as an anchor point.