If we presume that the AK-47 is a very well-designed firearm, and 50+ years of its use would seem to indicate that it is, then your response incorrectly presume

If we presume that the AK-47 is a very well-designed firearm, and 50+ years of its use would seem to indicate that it is, then your response incorrectly presumes that the person who acted with volition would have been equally successful when using a firearm of an inferior design. A death that occurred solely because the weapon allowed a kill by an otherwise insufficiently-skilled marksman can be attributed to the gun rather than its firer. The number of deaths attributable to the AK-47 would then be the number of total people killed by people using AK-47s minus the number of people who would still have been killed with the best available alternative weapon. It's a much smaller number, but it's a positive number none the less.