It's a 24-episode adaptation of an SF novel. (28 episodes if you count the "intermissions".) It carries its length thoughtfully and interestingly, but I don't imagine I'll rewatch it. It's better as an exploration of its ideas than it is on emotion and character.
It's set in a rather bland future. Theoretically we're 100 years ahead, but the differences between then and now are minimal and fairly cosmetic. The only exception is androids, aka. humanoid Interface Elements, or hIEs. They're everywhere. They look and sound like people, but the story's very, very clear that they're just algorithms running on software. They've got no soul or heart. They can't do anything without an owner. (There are also super-intelligent AIs, but those are kept isolated. Mankind is paranoid about what they could do, which is understandable since they have enough predictive power that they can functionally see the future.)
This series is mostly interested in exploring that. It's also got guns, action, military operations and so on, but we're never far away from some pretty good philosophising. What do people think of hIEs? Are we superficial idiots who react to them as if they're people, e.g. falling in love with one simply because it's pretty? (There's one of those and he's the hero.) Are we cavemen who resent and hate the hIEs much in the way that previous generations would bash immigrants, globalisation and ethnic minorities? (There's one of those too.) Alternatively, are we the kind of fanatic who'd sooner kill his friends than endanger the human race by trusting hIEs and AIs? "Because she's far more dangerous than any nuclear weapon."
That last one's particularly dangerous, incidentally, because his logic's not wrong. The problem's simply that his attitudes and priorities are themselves worsening the situation. (We're talking about the kind of person who'll get snippy with you for being polite, as if he thinks being polite to hIEs is an insult to all humans.) A super-intelligent AI could indeed trash world civilisation in half an hour and indeed has probably already devised a number of scenarios as a prediction exercise.
There's some pretty cool speculation here. How do you define a human? (A reasonable case is made for an unintuitive answer to that question.) The hIEs' dependence on their human owners is explored and even justified. (There's a class of high-level decision-making problems that are tough for a machine, if only since it can be fiendishly hard to frame the problem accurately. Delegating those decisions to its owner frees up an AI to focus on what it does well.) What if you put robots in charge of politics? Robot mayors? Entire robot cities, even? Different people have different ideas of what they believe the future should be. It's exploring trust, which is a fundamental element in human-robot relationships and a lack of it is likely to lead to bitter men going around in gangs at night attacking hIEs. The trust between Arato and Lucia is arguably their greatest weapon, capable of making them a dangerously powerful combination. There's also the concept of an "analog hack", i.e. when human-looking machines exploit psychology to elicit a specific emotional or behavioural response.
Oh, and robot sexual behaviour is only unlocked for users aged 18+.
Unfortunately I don't think the story always supports its level of thematic focus. I'm thinking of Kouka. The story gets really into choosing or designing the future. What future do you want? What future does he/she want? Can an hIE want a specific future at all? That's okay, as far as it goes, but for a long time it feels abstract. It's not always clear why some of the characters are focused on that, beyond the fact that that's what the author's writing about and the cast are subservient to the story. That's particularly true of a rampaging idiot psycho robot with the emotional intelligence of a six-year-old. She makes some extremely odd decisions. Eps.15-16 didn't work for me, personally, because her motivation felt too abstract and unintuitive.
Also, more trivially, the show's banging on about designing the future when it hasn't put much effort into designing its own future. Bar a handful of gadgets, it's basically what you'll see outside your window now. They have self-driving taxis. That kind of thing.
I quite like the show's mild dissection of its protagonist, Arato. He, his friends and the show itself are all very aware that he's superficial. (Even his airhead sister calls him an idiot.) He's thoroughly polite and nice, but his reactions to hIEs are due to how they look. He sees what looks like a person, so he's polite to what looks like a person. That's as deep as his reasoning goes, although personally I don't see what's wrong with politeness. (I note that the friend who objects the most to that is rude to everyone, including humans.)
That said, though, Arato's also a bit bland. He's the stereotypical Nice And Slightly Wet Anime Protagonist. Lucia's not much more distinctive, despite having robot superpowers. You like them, but they don't often make you laugh or anything. Lucia's robot sisters can be terrifying, but basically this is a show where characterisation is at the service of the arguments being explored. Everyone's there to voice a specific viewpoint.
I enjoyed the show. I charged through it, for what it's worth, and it stood up well to being marathoned. It's thoughtful, while also having a walking apocalypse called Snowdrop who can subvert all technology (in the 22nd century!) and a one-woman death machine called Methode. Even for Lacia, facing Methode in hand-to-hand combat is suicide. She's also capable of shooting an energy blast that's more like an erupting volcano. It can portray super-intelligent AIs convincingly. There's a girl who hates the world. I've been mildly critical of the characterisation, but it was good enough to carry 28 episodes and care about Arato and Lacia.
"If you see them as equal to humans, you don't value humans enough!"