I dug up my AAAI reviews from two years ago, and it looks like they've been tinkering with their scoring system, so I can't tell you a precise answer. I can give the general principle, however:
EasyChair allows the chairs of a conference to configure a set of "attributes" for scoring. Each attribute gets both a numerical value and a descriptor. Some examples from AAAI two years ago:
5: (ground-breaking (top 15%), a sufficient basis for accepting the paper)
4: (positive, a factor in accepting the paper)
2: (problematic, a factor in rejecting the paper)
2: (Strong acceptance. A 5 or 6 in some category, no 1 in any category.)
-2: (Strong rejection. A 1 in some category, no 5 or 6 in any category.)
So as you can see, then they were using a 1-6 scale and a -2 to 2 scale.
From the sample you gave, my guess is that they're currently using a 1-10 scale, with 5 being neutral. Thus
Would translate approximately to "8 out of 10 (a very good score, three positive steps above neutral)."