Mine is better than yours here as well. I wonder if I'm lucky, you're unlucky... On uA AC, mine was spot-on when compared with a Fluke 87V.
Perhaps I rate current measurement accuracy in a different way. Putting two meters in series makes the limitations of both meters influence the other.
I connect a 0.01% accuracy resistor in series with a voltage measured with 0.0015% accuracy and measure the current with a single meter to see how that compares with the current based on Ohm's Law. That's what I call real world current. Good meters are accurate without any correction required. A current to voltage converter using a ADA4530-1 IC which I made doesn't need any correction. I have tested it on measuring the current from a 1,000 volt power supply without having any problems.
I realize that $20 meters can't meet that standard. However, we are at the point where maybe a $500, definitely $1,000+ meters should. After making the above measurement, I correct it by taking the meter's shunt resistance into account. That produces a figure which should at least meet the manufacturer's specifications. But, if the meter is measuring a current which doesn't have a linear voltage to current relationship, it may be impossible to translate a measured current to the actual one. The same is true if accurate voltage regulation is required.
While all of this isn't that important for a $20 meter, I still evaluate equipment the same way. The manufacturers will keep making meters with compromised accuracy until meter reviews start measuring real world current which doesn't require you get out your slide rule to translate.
Back to the AN8008--while a 2 uA AC current measurement with reasonable accuracy is possible, 2 mA is not.