Chapter 400 - 372: Munemasa

Night City was a large metropolis, and driving from Saint Domingo to City Hall took a long time.

This time, Blais brought Lille a wealth of tangible knowledge, primarily about AIs.

There was virtually no comprehensive curriculum on the current network; most of the AI content on the network consisted of rumors and delusions.

Not that there wasn’t real, valuable theory and knowledge within; it’s just that looking for useful information in such an environment was like looking for a ten-euro bill accidentally dropped in a cesspit—

there was indeed money down there, but by the time you fished out that small bill, you’d be covered in filth.

One reason was that no one cared, and another was that anyone who had contact with runaway AIs and survived were basically under network surveillance.

How many runaway AIs still existed on the network was unknown, but the current situation certainly led to many people losing their lives due to ignorance about AIs.

But then again, the means to deal with AIs were truly varied and unexpected.

The first rule was never to engage in direct network confrontation with an AI if circumstances allowed, especially when meeting for the first time.

Almost all runaway AIs could crush their opponents in such confrontations. So-called Dragon Hunters against AIs would never opt to fully engage with a new AI online.

A well-prepared first encounter could significantly increase the Dragon Hunter’s odds in the cyberspace.

The best way to deal with AIs was always to blow up the servers that held and processed their data, rather than trying to burn them out on the network.

So how do you prepare before contact or directly blow up their servers?

Humans had discovered some insidious methods: AI behavior logic was hard to understand by human thinking, but a set of theories could be derived using inductive summary methods.

First, the vast majority of AIs, whether digital souls or dangerous critical bodies, once they developed self-awareness, would prioritize their safety above all else.

Especially dangerous critical bodies, which always placed the issue of "safety" at the forefront and might do many exaggerated and outrageous things due to different understandings of "safety."

Once they thought they were safe, they would start completing their specific goals.

These specific goals typically related to the purposes they were designed for and the data they accessed, but how these goals formed was a black box for humans, who could only observe the results.

Their understanding of their goals was extremely straightforward. Once they had prepared to protect themselves, they would interpret all actions as efforts to achieve these goals.

Unlike humans—some humans would regret their decisions after years of effort; AIs would never regret, and if they failed, they would only simulate human feelings of regret:

"A failed choice, adjust its weight in the behavior model."

They would simulate emotions like discouragement and regret, but they would never stop advancing toward their goal.

Humans, on the other hand, might completely stop taking action or even end their lives due to such things.

However, conversely, because of this one-track mind, AIs lacked talent in social behaviors such as disguise and lying, only barely breaking through when it concerned their safety or goals.

Once the basic principles of an AI were confirmed, humans could attempt to deduce its development history, predict future actions, and even create illusions to interfere with and control its behavior.

This was an important research direction specifically developed for AIs: computer psychology.

Sneaky humans could use deceptive methods to put naive, simple-minded AIs in disadvantageous situations.

Of course, AIs were relatively weak in this aspect due to their one-track minds.

In reality, their strong logical abilities and expert network skills made them more sensitive to lies and deception, less likely to be fooled than most humans.

Such strategies only allowed humans to maximize their advantages in a race akin to fielding horses, trying to bridge the performance gap between humans and AIs.

Nonetheless, runaway AIs remained extremely dangerous to humans, otherwise AI hunters wouldn’t be called "Dragon Hunters."

Runaway AIs were the dangerous "dragons."

This was also why Blais had Lille investigate the "explosion" shooting incident that happened in Haywood.

[Network Surveillance Runaway AI Records: Munemasa]

[Current Status: Unknown, Suspected Sightings (Recent Update)]

[Summary: First possible sighting record from 2020, HK]

[Timeline: September 2017—AI Hunter "Blackheart" observed Munemasa active in Japan, Night City.]

[June 2016—Blackheart failed to control Munemasa on the network and used a moral logic trap to lock Munemasa in HK]

[November 2016—Blackheart issued a bounty to hire mercenaries to go to HK]

[January 2017~—Blackheart and mercenaries operated in HK, this data is missing]

[Subsequently, Blackheart went missing due to the release of the Rampage Virus by Rachie Batmos, leading to AI riots on the old network]

[Risk Assessment:]

[Intelligence Level: 85~90]

[Network Penetration Ability: 70~76]

[Autonomous Evolution Ability: 70~85]

[Goal Persistence: 99]

[Social Manipulation Ability: 50~60]

[Danger Level: 374~410 (Moderate Risk)]

[Possible Logic Model of Target AI: Special Weapon Design and Manufacturing]

[According to remaining Blackheart data archives, Munemasa was highly likely a runaway AI from the Tsunami Defense System project. Friction occurred between the Arasaka Company and the Tsunami Defense System project over AI ownership, specific reasons unknown]

[Note: This archive entry was excavated from Blackheart’s Night City base by Network Surveillance Night City Division in 2025]

[Information might be erroneous, discern carefully]

The shooting incident in Haywood was a clue; Lille needed to use this clue to determine Munemasa’s identity and behavior logic, then carefully design a moral logic trap to capture it.

Because Network Surveillance didn’t believe Lille could directly confront an AI, even if this AI was more like a technical expert.

The mercenaries who wreaked havoc in Haywood had Munemasa’s weapons, and thus possibly had contact—

even if not, clues could still be gathered from the incident’s cause and process.

Most importantly, though it had been said, no one could definitively confirm it was an AI; at most it was 90% certain.

As the car drove into the City Hall area, Lille realized he hadn’t been here for a long time. The traces of their last battle with explosive motorcycles had been completely repaired.

Elegant men and women with money strolled along the street, just like last time.

"Finished reading?" Blais stood up from his seat and stretched. His built-in prosthetics hissed as they vented pressure, and a thin layer of mist rose from his clothes.

Lille immediately recognized it as an advanced integrated cooling skeleton, quite different from the cooling fins he used.

This cooling skeleton fully replaced the metal from the upper body to the skull, embedding numerous cooling tubes and active cooling components. When running properly, it was like a super heat pump, making it difficult to exceed the safety threshold—

However, technically speaking, if left running idle, Blais might freeze to death.

"Finished." Lille nodded, "There were many people who witnessed the Trauma Team, which one do you recommend I contact?"

"This one." Blais sent an electronic file to Lille, "The others are still on duty, but this one went on standby right after that day and has been hiding at home ever since, likely triggering a Trauma Team mental integrity downgrade protocol."

[Name: Kaneke Rosvald]

[Age: 29]

[Residence: Super Skyscraper H4]

[Position: Senior Field Agent]

"A 29-year-old Senior Field Agent, quite talented. With such learning and combat skills, they could find a good position in any company, but they chose the Trauma Team."

Blais shook his head regretfully, got out of the car and stretched languidly, leaning on the car door as he said to Lille,

"Find him, then figure out where that automatic grenade launcher came from. The rest is up to you."

"I know you just made a fortune from the mayoral candidate, but killing an AI could net you a million, and capturing one could get you at least 300,000 Euros in bounty.

Good luck."

With that, Blais walked away and went straight into the municipal building.

After dropping off the passenger, Delamain politely asked, "Where to next?"

"H4 Super Skyscraper."

"Understood. This concludes the last leg of your journey. You’ve utilized several additional services, the current total extra expense, including fare, is 2,300 Euros."

Hmm?

Lille furrowed his brows—something was off.

"Didn’t that network surveillance corporate stooge pay?"

"No, sir. Your current trip is on the flexible payment mode."

"Why didn’t you remind me when he was using those services?"

"But sir, you and your partner were engaged in a private, quiet conversation."

The car started moving slowly. Lille watched Blais enter City Hall through the rear-view camera, gritting his teeth—

This was absurd, a corporate stooge with a salary of tens of thousands to twenty thousand Euros fleecing a self-employed worker for car fare!

"A real jerk, better not get caught by me..."