The case for Armor Units

[hcobb]

The Pentagon's Third Offset is retaining "human in the loop" for kill decisions, and it's a good thing.

Take the F-35 (please). How can the programmers judge years ahead of time which faint stealth blips are actual hostiles, which are decoys and which are civilians or incapacitated friendlies? It requires human level judgment (which is constantly only a few decades in the future of course), because the cost (in ammo alone!) of firing automatically at each blip is too high.

Hence if your human is distracted you'll pilot the vehicle to avoid running into anything while firing only in point defense. I.e. the GEV "D" result.

[ianargent]

The case against :judgemental" AI will never be a technical one - the advancements in AI that have allowed for the most recent level of skill for Google Translate, and the various voice-activated personal assistants, makes me somewhat leery of betting against Moore's Law.

Instead, it's a question of responsibility. Who hold the "responsibility" when an AI makes a mistake? It will take a major societal shift for responsibility to be assigned away from a human, and in warfare, nothing is certain enough for any human to want to give up control.