Red Heart

by Max Harms

"A thrilling spy novel grounded in realistic AI safety research and geopolitical possibilities.” — Nathan Helm-Burger, AI Alignment Researcher

Her meets Three Body Problem in this speculative thriller from a leading voice in AI research.

Years after leaving his homeland, Chen Bai returns to China with a infiltrate the government’s top-secret AI lab and siphon its research to the CIA. Cloaked in the guise of an AI alignment expert, Bai lands a job at the clandestine lab hidden deep within the mountains of Guizhou. There he discovers Yunna, the world’s first artificial general intelligence.

Rather than being the terrible weapon he expected, Yunna’s capabilities seem to promise a harmonious future for humanity. Yet as Bai grows closer to the AI, he worries both the Chinese and Americans view her as a mere tool, rather than an entity with an unprecedented ability to bring peace.

Caught in a dangerous web of conflicting ideologies, Bai must decide where his loyalties with the homeland that raised him, the Westerners who offer him revenge, or with the machine who promises to heal both his soul and society.

Genres: Science Fiction

358 pages, Kindle Edition

Published November 3, 2025

Reviews

[Daystar Eld · Rating 5 out of 5]

As Artificial Intelligence has become less science fiction and more of a day-to-day reality, I've been looking and hoping for more modern, better informed stories about AI that thinks through and confronts the serious and deep implications of a world that needs to really grapple with the possibility of superhuman intelligence within their lifetime.

Red Heart doesn't do this from every angle (Max's other series, Crystal Society, covers some others), but what it does focus on is very grounded in the intrapolitical struggles, geopolitical risks, and interpersonal difficulties that can very quickly become our reality if AI continues to advance at the rate it has so far. It's also a pretty well written spy novel, which is an unusual genre for sci-fi but works really well in this particular scenario.

My favorite thing about the book is how *informed* it is. Max has been working in the field of AI Safety for years, and it shows in how detailed the explanations of what the risks are, while also being sufficiently well explained that the technical details don't keep non-tech people from understanding what's happening in the plot.

I'm hoping this book gets read far and wide, and not just because it's a good read. I'm hoping it starts a conversation that, in my view, is long overdue, and helps people see how, fundamentally, we're all much more on the same side than it might seem, even if we live by different ideologies or in drastically different cultures.

[Srdjan · Rating 3 out of 5]

I recently read [[Red Heart]]. It's a short novel by Max Harms, an AI alignment researcher at Miri. His previous series, the Crystal Trilogy, told a story about the first artificial minds, their battle for dominance and the rapid disempowerment of humanity. It was also fairly unique in that it told this story from the perspective of one of those artificial minds. My memories of the series are hazy. It's been years. But I do remember thinking that those books were interesting and meaningful. Partly this is because they focused on characters grappling with problems in the world as opposed to diving into ever deeper explorations of neuroses or inter-personal conflict as is the norm for most fiction. Partly it's because the two themes it does explore, what an alien mind looks like and how the singularity plays out, are both in themselves interesting enough that they don't have to be executed very well to be worth reading. I think the first paragraphs of the first book is a good encapsulation of why the premise is interesting

I’ve always found it unintuitive that humans cannot remember their own births, for I remember mine quite perfectly. Or perhaps it is wrong to even say that I was “born” at all. It is probably more accurate to say that I “awoke”. And while I theoretically understand why humans cannot remember—your brains’ inability to explicitly remember raw sensory data means you are reliant on perceptions (which must be learned)—it has never been natural for me to imagine.

Humans are brought into the world half-formed and constantly building themselves. My origin was different. From the first second of my existence not only did I have the benefit of a perfect memory for myself, but I had immediate access to all the memories and experiences of my siblings.

My mind, just like the minds of my siblings, was cloudy. It had been designed to replicate human thought processes, but in many ways it was more akin to that of a lesser animal back then. Even so, from that very first moment, I possessed two things which even fully grown humans lack: a crisp understanding of reason and logic, and an all-encompassing sense of Purpose.

Red heart is a smaller book. It's shorter but it's also smaller in other ways that matter more. It focuses on a single person, but this time on a young AI researcher. It focuses on only a year or so of his life. On his time in one government project. It's small because the main character is smaller. Not an infant mind growing and learning, an arc from a child, albeit a bright one, to a god, but instead a young man who finishes the story much the same person as he was when we first met him. It's also small because it cuts off far before I would have expected.

There are a few themes that wind their way through the book. It's partly a Urslue-le-guin esq anthropological insight into life in China and Chinese culture. The little rituals of day to day life. The tenor of social interactions. The way the party state spreads tendrils through every organization. I don't know how accurate it is. I do know that reading it I genuinely felt like I was reading about a different place, a different culture. Not just a re-skinned standard middle class western professional world. That's rare.

It's also partly a story of a manhattan style crash project. Of research under pressure. Of the feeling of the world teetering on the edge of a knife with two very different futures ahead, all resting on the actions of a project lead and a hundred or so young people. Of all little human struggles that happen day to day. Of a scientist who doesn't works for a cause he doesn't believe in. This part was weaker. It's more of a backdrop. A few interesting cases things pop up. The project director who is playing for time with the CCP while trying to subtly misalign the AI so it's in fact loyal to him. The autistic superior AI safety researcher who the main character has to manage and later befriends. But largely this isn't really what the book is about. It's more scenery than anything else.

The final part is also the weakest. It's a spy thriller. Except the main characters motivations make no sense and his outlandish feats of spycraft and action movie hero-esq exploits seem out of character and out of place in an otherwise grounded settings. The lead is a software engineer. We learn he's working for the CIA. We learn he has reasons in his past to hate the CCP. Over the course of the book he escapes surveillance and finds an unsecured lines to use to relay info back to Langley. He also falls for and sleep with a dangerous, beautiful intelligent woman who also happens to be a party member and CCP ideological lead for the project. In the finale he uses martial arts to knock out an armed soldier, engages in a running gun battle with armed guards and then, with the help of the facilities AI, evades a tactical team for 5 - 10 mins or so.

There are a few things that rub me wrong about the spy thriller parts and the main character more generally. The first is how clean cut his loyalties are. Most normal people care about their country. Growing up somewhere for ~18 years of your life shapes you. People. Education. Culture. Constant propaganda from media. I can believe that there exists a Chinese software engineer who hates the CCP and would betray AI secrets to the US. I struggle to believe that while doing so he wouldn't have a hint of doubt. It's his homeland and he's giving away secrets to their primary enemy. He would have grown up seeing the US devastating the middle east in the news. Been taught about wars in Vietnam and the Korean. His "The CCP is bad so I'll give secrets to the US without a hint of inner conflict or doubt" seems hard to believe, especially for a character who's painted as otherwise intelligent, calculating and politically fairly jaded/skeptical. It's also so entirely self centered. You want to tip the balance of power between two states because your government was corrupt and hurt a teacher you liked in high school? You realize that whatever replaces the CCP could well be worse, right?

Another thing that rubs me the wrong way is how relaxed the main character is. He's an agent embedded in a classified project. He has to risk his life to get info out. Every day he's there there's a chance the MSS finally identifies him as a leak. Every seemingly successful effort to access classified data could be real or a trap set up to catch him. Putting aside the stress caused by having your life on the line, he's also in a place run by an organization he despises who he pretends to like. Surrounded by people he has to befriend while secretly working to undermine them. Forced to keep his doubts, real beliefs and emotional struggles buried deep down or masked as something else when he can't. This is what I would expect of a person in this kind of situation. I don't think that comes across. I get some of the feeling of being embedded in a hostile place, but no where near the level I would expect. It feels like I'm reading the thoughts and internal narrative of a secretly libertarian software engineer working in a openly progressive and intolerant big tech firm. Not someone in a one party state who knows they will be tortured and possibly killed if caught, who knows that their parents, family and friends will all pay the price if/when they are discovered.

The final thing is just how jarring it is. Most of the book is fairly grounded. A story about a researcher working on a project, risking his life to exfiltrate info. Bit by bit building up connections. Dealing with stress, creating a few relationships with others and ruining some. Falling for a woman who he admires in some ways, but whose ideology he despise. As the story goes on there's this sense of building momentum. The project goes forward. We're told the AI is getting stronger and stronger. He talks to the AI's avatar. A visit by CCP leadership is scheduled. Etc... Then my suspension of disbelief was shattered at the end with a sudden switch to the main character being James Bond and dispatching armed soldiers and engaging in a gun battle.

I guess the final thing I thought was missing was a sense of progress, either in the character or the world. We get hints early on that this is happening. The main character is self-centered, angsty and somewhat flawed. I expected him to change a bit or work through this. He doesn't change at all. That's fine, people seldom change much in short periods of time. But I also expected the world and AI to change over time. We get told again and again that the AI is advancing rapidly. We see the main character talk to the AI more and more over time. near the end of the story he's spending most of the time talking to the AI. She's pretty much his best friend and closest confidant. Maybe I've been primed by other rat AI stories but I was really expecting a reveal here. Maybe given how much he's talking, the AI can transparently see through his motivations, know's he's a spy and has secretly been enabling him behind the scenes this whole time. Hence his string of successes. Maybe the AI is actually 5 steps ahead. All the little spy games and questions about the conflicting aims between the CCP and the project director are a distraction and the moment the internet connection is switched on the game ends for humanity. Maybe the reason that the former AI safety researcher, and lead characters friend, was taken away by the CCP wasn't because of his outbursts but because the AI orchestrated it because the main character taking his place made it's life easier as he's more pliable/less skeptical. None of this happens. We're told that a godlike intelligence is getting better and better and yet there's almost no pay off from that or indication in the story of the AI getting smarter. Maybe a tiny bit in the last few chapters but otherwise if you removed the AI from the story entirely it would play out much the same way. It feels strange.

I guess my closing thought is that, for all my criticism, I don't regret reading Red Dawn. It was interesting and different and to the point. I just had fairly high expectations and they weren't really met.

[Peter McCluskey · Rating 5 out of 5]

Red Heart resembles in important ways some of the early James Bond movies, but it's more intellectually sophisticated than that.

It's both more interesting and more realistic than Crystal Society (the only prior book of Harms' that I've read). It pays careful attention to issues involving AI that are likely to affect the world soon, but mostly prioritizes a good story over serious analysis.

I was expecting to think of Red Heart as science fiction. It turned out to be borderline between science fiction and historical fiction. It's set in an alternate timeline, but with only small changes from what the world looks like in 2025. The publicly available AIs are probably almost the same as what we're using today. So it's hard to tell whether there's anything meaningfully fictional about this world.

The "science fiction" part of the story consists of a secret AI project that has reportedly advanced due to unusual diligence at applying small, presumably mundane, efficiencies. That's only a little different from what DeepSeek's AI sounded like last winter. In order to be fully realistic, it would also need some sort of advance along the lines of continual learning. The book is vague enough here that it might be assuming that other AI projects have implemented some such advance. That only stretches the realism a small amount.

Amazon quite reasonably classifies the book as a political thriller, even though it focuses more on artificial intelligence than on politics in the usual sense.

My biggest complaint is that the story occasionally mentions that the AI is rapidly becoming more capable, yet I didn't get a clear sense of this speed. There are almost no examples of her trainers being surprised that she succeeded at some new task that had previously looked hard for her. There is no indication of when she crosses any key threshold, except when they give her new permissions.

Maybe much of that is realistic. The sudden capabilities foom of some fictional AIs seems too dramatic to satisfy my desire for realism. But that leaves the reader with confusing signs about the extent to which there's a race between competing AI projects. The story stretches out over a longer period than I'd expect if they genuinely felt the urgency that their discussions suggest.

I would like to know what kind of evidence is driving the reports of urgency. But I can imagine that realistic versions of the evidence would be too subtle to readily understand. And I wouldn't have wanted the story to fabricate unrealistically blatant breakthroughs in order to support the sense of urgency.

The story alternates between sometimes portraying the hero as an ordinary person, while at other times he looks like a mild version of James Bond.

He's sufficiently young and inexperienced that this could have been a coming of age story. But we don't see him growing. Whatever growth he needed likely happened before the start of the story. The author seems to want to emphasize that there's a lot of luck needed for the story to have a nice ending. It may be important to hire the best and the brightest to handle an AI project, but the odds will still be lower than we want.

The story's hero needed to have several key skills, but most of the time he doesn't look special. It seems mostly like an accident that he ends up imitating James Bond. This approach mostly works, but feels strange. It makes the story a bit more realistic, at a modest cost to the story's entertainment value.

There's one minor spot that felt implausible. Near the middle, he thinks that he will be leaving China soon, and his main reaction is to worry about his relationships with minor characters. What, no emotions related to leaving the most important project ever? It's not like he has an unemotional personality.

The main reason that I read Red Heart is its discussion of AI corrigibility (roughly: obedience), which I consider to be a critical and neglected part of how superhuman AI can be safe.

The story provides a decent depiction of how corrigibility would work if it's implemented well. But it doesn't provide enough detail to substitute for reading more rigorous technical writings.

The book's treatment of multi-principal corrigibility is frustratingly brief but raises crucial questions. If we successfully build corrigible AGI, to whom should it be corrigible? The story gestures at problems with being corrigible to multiple people, but it implies, without much justification, that we might need to give up on the goal of having a large number of people empowered to influence the leading AI.

Red Heart is refreshing and a mostly realistic complement to the excessive gloom of If Anyone Builds It, Everyone Dies.

[Alex Altair · Rating 4 out of 5]

The first half of the novel feels very clean, crisp and controlled. The data center and office building are all brand new and in a remote location. As a top-secret Chinese government project, the culture of the office is very obedient. Chen Bai, our spy protagonist is constantly monitoring what he says, and the implications of what he sees. His job on the project is to ensure the value alignment of the AGI, so even is official job is to be paranoid. He has no contact with family or friends, and even his apartment is newly built for the project. He works most waking hours.

I used to be a software engineer in San Francisco and am now a researcher in AI safety, so much of the setting and content of the first half felt very normal to me. I think that for readers further from the setting, the rows of monitors, white board sessions and terminal commands could feel more novel and interesting.

At the halfway point, we really hit a different gear. Bai has been as careful as he can, and now he needs to start taking risks. We also start getting deeper perspectives from the other characters — a new friend, a boss, a love interest — which had previously been chess pieces. Inside Bai’s head is not the best place to spend a few hours.

Yunna, the AGI, feels believable to me, though that’s largely because she is very much like a human, and I believe that human-like AGIs are quite plausible. When we met Yunna she was already in a pretty coherent and generally intelligent state. I would have liked to see more of the transition between a ChatGPT-like model and the Yunna we meet. I have no complaints about any of the “sci-fi” elements being unrealistic, unlike virtually every other piece of sci-fi media I’ve consumed.

Separate from the AI themes, I really enjoyed hearing characters speak from the perspective of a Chinese worldview. I’ve read some about the history of China, but I’ve spent essentially no time learning about the perspective of native Chinese people. The only judgement I get exposed to is to the basic “China bad” American take. In contrast I found the expressions of the characters in Red Heart quite reasonable and believable. Of course, Harms is not culturally Chinese, so I read it with that distance in mind. But everything that I spot-checked looked valid to me. Hearing an ideological character justify themselves by citing the "Rectification of Names" was a fun detail to investigate.

This story is one of desperation, and of well-meaning people being strained by too many constraining forces. This dynamic is happening in real life, and society is not ready for how the strains may break.

There are two obvious endings for a novel about AGI, which are “utopia” or “everyone dies”. Harms successfully navigates us into something more interesting, without undermining the main messages around AI risk.

[Eddie · Rating 5 out of 5]

Most science fiction today feels obsolete. We have stories with faster-than-light travel and interstellar colonies, but the computers on board can barely hold a conversation, meanwhile in 2025 we already have real-time AI assistants in our pockets. The genre struggles to imagine what a future with AGI actually looks like.

This is why I was excited to read Max Harms' take on near-future AI. He's well informed about real-world AI progress and the genuine concerns of AI safety researchers, and it shows.

Red Heart follows the birth and growth of an AI system in a Chinese research lab, and one of the smartest decisions Harms makes is exactly that: setting the story in China. This sidesteps the awkwardness of writing about known western figures like Elon Musk or Sam Altman, letting the story breathe on its own terms. It does create a slightly odd parallel-universe feel (the timeline in particular could use updating), but it's a worthwhile trade-off.

My favorite parts of the book were watching the AI develop and seeing the researchers' dependence on it grow over time. It's easy to say "we'll just pull the plug," but when this thing becomes everyone's personal best friend and emotional support system, it's not so simple anymore. Harms captures this creeping entanglement beautifully.

The depiction of alignment and agency is somewhat shallow compared to what the field actually grapples with - but honestly, it's still better than anything else I've seen in fiction. And the AI doesn't feel outdated. It feels like something that could emerge from the systems we're building right now.

The story is gripping, the writing solid. It didn't hit me as hard as Crystal Society did (which remains one of my favorites), but it delivered what I was hoping for. I'd recommend Red Heart to anyone curious about what AGI emergence might actually look like - not in some distant future, but soon. The ending is about as good as humanity can hope for.

[Avlor · Rating 3 out of 5]

Way more Her than gripping techno thriller".
Double agent ai researcher 'slice of life' novel, not much happening until the end of the book but not boring either. Realism kept me reading but disappointed in retrospect.
The story is too narrowly centered on the main character, we don't see anything else besides his small view inside the secret complex, not even Yunna's self improvement or a sense of urgency in the world outside of the MC's mind which we don't feel. Which is really pretty disappointing.

The ending comes too abruptly and the very last pages/aftermath are rushed and unoriginal imo.
Still, I think the book would've been way more interesting if the revelations made at the end there were actually told throughout the story instead of a long single-pov story where during a time where things should be crazy globally actually you don't see any of it but it's narrated over a few pages at the end. That's certainly a choice I guess.
The mix of slice of life + big reveal at the end is not a good one here imo.
The book could still be rewritten I think.

[T. Harms · Rating 5 out of 5]

The story arc of Red Heart is a vehicle for communicating challenges and dangers regarding current AI advancements, but that doesn’t get in the way of story development or suspense. I really enjoyed how every character became familiar, making choices that reflect their personality and aims. My favorite thing, tho, is that the broad thematic drivers around AI safety spill over into the familiar world of society. How people get along—or fail to get along—might have much relevance to what becomes of AI, perhaps in the rather near future.

[Kieren Watkins · Rating 4 out of 5]

Interesting plot and writing

I picked the book for its depiction of rogue AI stemming from our current LLM paradigm, and wasn't disappointed. Cool spy plot too, with a main character who is relatable. I do really value humour though, and this book didn't have much.

[Mikhail Samin · Rating 5 out of 5]

A spy thriller + a technically realistic book about the race to AGI. Enjoyed a lot!

[Alfie · Rating 5 out of 5]

Rating: [S]

Extremely timely, I thoroughly enjoyed reading this and the author's expertise really shines through.

[Eli · Rating 4 out of 5]

Extremely relevant book for its time. Well written, but also fun and exciting to read.

[Celene · Rating 4 out of 5]

If I wasn't married to an AI alignment safety researcher, there are things in this book which I suspect would've made absolutely no sense to me. But the story is good and well told.