Robotic exposition about robots this way.
So, this actually is the end of the arc, and consists primarily of a long, draggy series of speeches from the mad scientist and her now criminal old flame about the RF models (of which, recall, Harold is the only surviving example). Evidently, they were made with a system based on human brain maps, giving them inner worlds far vaster than that of normal robots and, perhaps, the ability to form legitimate thoughts of their own.
There, that didn’t take long to say. Why does it take the show half its running time to beat that point into the ground?
Anyway, eventually this ends up with a scuffle over a tablet set to upload Harold’s real core code, which would cause him to be executed if it got out. After standing there like an idiot for most of the episode and getting punched down by a nerd with embarrassing ease, Echika gets the tablet and pauses the upload. By this point, mad scientist Lexie has been shot in the leg and the real culprit through the lower torso, so being the last person in the room without crippling bullet holes, Echika gets to decide what to do. She pulls off some mental gymnastics to not send Harold to a nice farm upstate, and all’s well that ends well with two arrested scientists and a cleared super-robot who wants to be more human and get closer to his partner. This also takes way longer than it deserved, which is a little rough.
Now, there are a couple of good points here. The biggest one is what should be a bombshell that Lexie drops which is buried in how much over-explaining is in this episode on the whole: the “Laws of Respect” that are this setting’s version of Asimov’s Three Laws of Robotics are actually a technical fiction: normal robots aren’t even smart enough to have such laws applied, they just lack the concepts of such forbidden actions. Harold is implied to be much smarter bound for different reason: the laws have never been hard coded because they can’t be, and in his case he’s limited because he believes himself limited.
Unfortunately, by the time this is delivered there’s been so much jabbering about black boxes and layers of lies that you’re probably half asleep. I actually love this kind of stuff most of the time. Some of the more interesting elements of Metallic Rouge were based around what it meant to be free or bound, Time of Eve and Gene of AI both dealt with AI ethics in intelligent ways, and Beatless had a wonderful discussion about how the Three Laws couldn’t be coded in their setting because of the fuzzy definitions of such terms. Not to mention that many of the classic Asimov stories are all about hiccups in the Laws, where they conflict with reality or the consequences of modifying them in any way. Even as someone who eats up the technical aspects of robot scifi and was thus able to follow it in this, the delivery made it a bore.
It seems, though, like we are really, finally, at an arc close. Sure, this could be a fakeout like episode 2, but I’m going to guess that no, this is where one of the light novels ended. So we’ll see next week if we can get a new plot to build on what we’ve got, hopefully with better pacing.