Chapter 372 Muramasa

Night City is a big city, and it takes a long time to drive from Santo Domingo to the municipal center.

This time, what Bryce brought to Lille was more substantive knowledge, mainly for AI.

There are almost no systematic teaching materials for this information on the Internet today, and most of the content about AI that can be found on the Internet are rumors and delusions.

It’s not that there are no real and effective theories and knowledge in it, but looking for useful things in such an environment is tantamount to looking for a 10-euro note that accidentally fell into a cesspit——

There is indeed money in it, but when you take out the small piece of money, the shit is all on you.

On the one hand, no one cares about this. On the other hand, in this era, those who have come into contact with circulating AI and are still alive are basically under network surveillance.

It is unknown how many AIs are circulating in the current network, but the current situation has definitely resulted in many people losing their lives inexplicably because they know nothing about AI.

But then again

This method of dealing with AI is really rich and unexpected.

The first rule of thumb is: never engage in direct online confrontation with an AI if allowed, especially if it's your first time meeting.

Almost all circulating AIs can crush their opponents in such a confrontation, and the so-called Dragon Hunters who fight against AI will not choose to fully output the AI for the first time on the Internet.

A well-prepared engagement can greatly increase a dragon hunter's chances of success in cyberspace.

And the best way to deal with AI is always to blow up the servers where they host and process data, rather than trying to burn them on the network.

So how to prepare before contact, or directly blow up their servers?

Smart humans have discovered some insidious methods: the behavioral logic of AI is difficult to understand with human thinking, but a set of theories can be obtained using summary and induction methods.

First of all, most AIs, whether they are digital souls or dangerous critical bodies, will put their own safety first as long as they have self-awareness.

Especially dangerous critical bodies, they always give top priority to the issue of "safety or not", and will do many exaggerated and outrageous things because of their different understanding of "safety".

Once they think it's safe, they begin to accomplish their specific goals.

These special goals are usually related to the purpose for which they are designed and the data they come into contact with, but how they are formed is a black box for humans, and researchers can only see the results.

They have an extremely direct understanding of their goals. Once they have completed their self-protection preparations, they will interpret all their actions as accomplishing this goal.

Unlike humans, who will regret the choices they make after years of hard work, AIs will never regret it. If they fail, they will only simulate the human emotion of regret:

“A failed decision, adjust its weight in the behavioral pattern.”

They will simulate some emotions, such as discouragement, regret, etc., but they will never stop and will always move forward step by step on the road to achieving their goals.

But human beings, human beings can completely stop moving forever or even end their lives because of these things.

But conversely, because this model is too rigid, AI lacks talent in social behaviors such as pretending and lying. It will only be forced to explode when these things are linked to its own safety or goals.

Once the basic concepts of these AIs are confirmed, humans can try to deduce their development history, predict possible future actions, and even create illusions to interfere and control their actions.

This is an important research direction specifically designed for AI: computer psychology.

Cunning humans can use deceptive methods to put honest and innocent AI into a disadvantageous environment.

Of course, AI is just one thing and is relatively weak in this regard.

In fact, their strong logical abilities and superb network technology still make them more sensitive to lies and scams than most people, and less likely to be deceived.

This kind of strategy only allows humans to use Tian Ji's horse racing method to maximize their advantages to fill the gap in comprehensive performance between humans and AI.

No matter what, rogue AI is still very dangerous to humans, otherwise AI hunters wouldn’t be called “dragon hunters.”

The wandering AI is the dangerous "dragon".

This is why Bryce asked Leal to investigate the "explosion" shooting that occurred in Haywood.

[Network prosecutorial circulation of AI file records: Muramasa]

[Current situation: unknown, suspected sighting (latest update)]

[Brief description: The first possible sighting record is from 2020, HK]

[Timeline: September 2017 - AI hunter "Black Heart" observed the AI's activities in Japan in Night City. 】

[June 2016 - Black Heart was unable to control Muramasa on the Internet and used moral logic traps to lock Muramasa to HK]

[November 2016 - Black Heart released a bounty to hire mercenaries to go to HK]

[January 2017~——Black Heart and mercenaries operate in HK, this data is missing]

[Subsequently, due to the release of the rabies virus by Lachie Butmos, the old network AI rioted, and the black heart lost contact]

【risk assessment:】

[Intelligence level: 85~90]

[Network Penetration Ability: 70~76]

[Autonomous evolution ability: 70~85]

[Target persistence: 99]

[Social operation ability: 50~60]

[Danger level: 374~410 (medium risk)]

[Possible logical patterns of target AI: design and manufacturing of special weapons]

[Based on the data files left behind by Black Heart, Muramasa is most likely a rogue AI that defected from the Tsunami Defense System. Arasaka Company and Tsunami Defense System had friction over the ownership of the AI. The specific reason is unknown]

[Note: This file record was excavated by the Network Supervision Branch in Night City in 2025 at the black-hearted Night City stronghold]

[The information may contain errors, please be careful to identify them]

The shooting that occurred in Haywood is a clue. Leal needs to determine Muramasa's identity and behavioral logic based on this clue, and then carefully design a moral logic trap to capture him.

Because network monitoring does not believe that Lear can confront an AI head-on, even if the AI is actually more like a technical expert.

The mercenaries who killed everyone in Heywood had Muramasa's weapons in their hands, and they might have his contact information——

Even if not, it is possible to get some answers from the cause and course of the incident.

The most important thing is, having said all this, no one can actually be sure that this is an AI. At most, it is only 90% sure.

The car drove into the Civic Center area. Lille had not been here for a long time. The traces of the last time they fought with exploding motorcycles here have been fully repaired.

Rich and fashionable men and women were strolling in the streets just like last time.

"Have you finished reading?" Bryce stood up from his seat and moved his body. His built-in prosthetic body made a pressure release sound, and a thin layer of mist was blown up on the surface of his clothes.

Leal could tell at a glance that this was an advanced integrated heat dissipation frame, which was different from the heat dissipation fins he used.

This heat dissipation skeleton completely replaces the metal material from the upper body to the skull, and embeds a large number of heat dissipation pipes and active heat dissipation parts. It really operates like a super heat pump, and the temperature is difficult to exceed the safety threshold——

But technically speaking, if the system were left to idle, Bryce would probably freeze to death.

"I've finished reading." Lear nodded, "There are many witnesses in the trauma team. Which one do you recommend I contact?"

"This." Bryce sent an electronic file to Lear. "Other people at the scene are still at work, but this person went into standby status after that day and has been hiding at home for the past few days. It seems that the trauma team has been triggered. Mental Integrity Downgrade Protocol.”

[Name: Kanek Rothwald]

【Age: 29】

[Residence: Super Skyscraper H4]

[Position: Senior On-site Operator]

"The 29-year-old senior operator can be said to be very talented. With such learning ability and fighting talent, he can find a good position in any company, but it is the trauma team."

Bryce shook his head regretfully, got out of the car, stretched out, leaned on the edge of the car door and said to Lear:

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

I know you just made a lot of money from the mayoral candidate, but if you kill an AI, you can get 1 million. If you capture it, you can get at least 300,000 euros in bonus.

Just work hard. "

After saying that, Bryce walked away and walked straight into the municipal building.

After seeing off the customer, Delamain asked politely: "Where should I go to my next destination?"

"H4 super skyscraper."

"Okay, this trip has entered the last segment. You have used multiple add-ons. The current additional expenditure including fare is 2,300 euros."

Um?

Lear frowned - something was wrong.

"The corporate dog that monitored the network just now didn't pay?"

"Yes, you are using flexible payment mode for this trip."

"Why didn't you remind him when he used those things!"

"But sir, you and your partner were in the middle of a private meeting that required silence."

The car started slowly, and Leal watched Bryce walk into the city hall through the rearview camera, gritting his teeth——