Inside the Dark Web Market for Japanese Data: Norton's Iskander Sanchez-Rola on Leaks, AI, and Defense

Open on YouTube ↗
Overview

In this interview, Iskander Sanchez-Rola, Senior Director of AI and Innovation at Norton, walks through live dark web sites where stolen data from Japanese companies is advertised, auctioned, and dumped. The central question is how exposed ordinary people and businesses in Japan really are, and what has changed now that AI is available to attackers. Sanchez-Rola's position is blunt: some of everyone's data is almost certainly already out there, Japan's old "language barrier" no longer offers protection, and the realistic goal is not deleting leaked data but knowing what has leaked and limiting the damage.

20 min read
0:53

Three Layers of the Internet

Sanchez-Rola began by separating the internet into three layers. The "surface web" is what people use every day, the part that search engines can index and anyone can reach. Beneath it is the "deep web," which simply means content that isn't publicly accessible, such as a company's internal database or accounts behind a password. The "dark web" is a smaller area that runs on a special network, with Tor as the best-known example. Reaching it requires dedicated software, and it can't be searched the way the normal internet can. The dark web is the layer the conversation focused on.

1:56

A Ransomware Group's Leak Site and Its Japanese Victims

The first site Sanchez-Rola showed looked the way many people picture the dark web: a terminal-style interface where commands can be typed. The group behind it uses a Japanese name, though Sanchez-Rola said they are not actually Japanese, or at least that is what researchers believe.

The site listed many Japanese companies. For a well-known high-end home appliance maker, the group appeared to have access to everything, and the data was, in Sanchez-Rola's words, "one click distance" from being downloaded. Another listing, for a very famous large Japanese corporation, showed a countdown timer. The group was demanding a large sum, and the timer had three days left. If the company paid, the listing would disappear; if it didn't, all of the company's data would be published. Sanchez-Rola stressed that this was happening in real time.

A third listing, for a famous Japanese gaming-related company, showed what happens after a refusal. The company had not paid, so the data was openly available. The group had even posted a comment saying the company knew it had been hacked, noted that the incident happened in June, and linked news coverage of the company's decision not to pay or even contact the attackers. Anyone who knows where to look, Sanchez-Rola said, can now download that company's data with one click, without paying anything.

Extortion First, Then Sale to Anyone

Asked whether money is the main motive, Sanchez-Rola said yes, but explained that the hacked company is only the first potential buyer. If it refuses, the group will sell the data to whoever pays. That could be a competitor engaged in espionage, or a foreign state hostile to Japan that wants large volumes of data about Japanese people. In his description, these groups don't care who they sell to as long as they are paid.

5:57

Why Ordinary Users Shouldn't Go Looking

The interviewer asked whether reaching these sites requires special software unavailable to ordinary people. Sanchez-Rola said the software is now quite accessible. The real barrier is knowing where to go. He compared it to an internet with no search engine: dark web addresses are long strings of random letters and numbers that nobody stumbles across or memorizes, and many are not publicly listed. Finding these sites takes specialized knowledge.

He repeatedly advised people without that knowledge not to try. Clicking around without understanding what you are doing can get your own machine compromised, and you may commit illegal acts without realizing it. If authorities find out, he warned, "they may knock on your door." Even if something is technically possible, there are consequences.

7:38

The Range of What's Exposed

As he scrolled, Sanchez-Rola showed how broad the victim list is. One entry was a school in Kanagawa whose data on students and teachers, roughly half a million files, was available for download. Other listings covered a karaoke business, izakaya, real estate operators, and internet cafes. On one listing the timer read zero: the deadline had passed five months earlier without payment, and the page had nearly 37,000 views, which Sanchez-Rola read as a sign that many people had accessed the data.

He then showed examples of what the data actually contains. A real estate and insurance company's leak included tenants' addresses, payment records, phone numbers, and identification documents. Another Japanese company's dump contained internal documents, Excel files with financials, computer data, and passports. One file showed a financial transaction complete with the company's seals, who signed it, and what was done. A hospital's leak included a floor map listing patients by name and location, along with medical records. A Japanese newspaper's listing showed what looked like direct access to a company computer, where documents could be opened and browsed. A famous Japanese watch company based in Tokyo had its internal documents exposed, which Sanchez-Rola said could include data on customers who bought watches as well as employees.

He insisted this was not a scene from a film but real data, and that people shouldn't assume their own information is absent: "it is." He added that this is not only happening in other countries; attacks on Japan are extensive and span many sectors.

When the interviewer asked whether attackers get only documents or the systems themselves, Sanchez-Rola said they have access to everything the company has on its systems: PDFs (including what appeared to be job applications), JPEGs, Excel files, and CSVs.

12:26

Prices: Cheap Cards, Expensive Bundles

Asked about going rates, Sanchez-Rola explained that sales come in two forms. Some items are sold individually and tend to be cheap. Bundles, meaning everything taken from a single company, are where the higher costs lie.

He showed a site selling Japanese credit cards specifically, with prices starting at about $2 per card, which he called pretty low. The seller listed more than 33,000 Japanese cards currently on offer and claimed a much larger total stock of more than 2.5 million. Buyers could filter by bank, card type, and the information attached. Sanchez-Rola noted that this was just one seller.

He also pointed out that not every site looks like a black screen with green text. One resembled an old-school video game while advertising data from multiple companies, including one offering around 8,000 driver's licenses. Another listing for a manufacturing and logistics company showed names, password hashes, IDs, emails, referral data, names in kanji, and employers, as well as internal server data. Sanchez-Rola emphasized that victims are not only consumer-facing businesses; logistics companies can be critical to the country.

14:31

The Disappearing Language Barrier

The interviewer asked whether anything is distinctive about Japanese data. Sanchez-Rola said there may once have been some truth to the idea that Japanese was a protective barrier, because it differs so much from English. But computers in Japanese companies "speak the same language, one and zero," so attackers could always breach them; what they struggled with was communication.

AI has changed that. Messages that once betrayed attackers through typos and awkward phrasing no longer do, and attackers can now target Japanese companies more effectively. He cited public data based on what hacker groups themselves report, not Norton's internal data: between January and October of this year, attacks on Japanese companies rose 40% compared with the same period the previous year. By his figures, that works out to roughly one hack of a Japanese company every two and a half days, versus every three and a half days the year before. Some groups focus on Japan more than others, but many target it, which he attributed to Japan being a large, important, wealthy country.

He also said attacks grow a bit each year and suggested AI may be one reason, since attacks that once required real expertise are getting easier. He was careful to add that this doesn't mean Japanese companies are doing a bad job; attacks are simply advancing.

There was a second effect of AI he highlighted. In the past, attackers sometimes didn't understand the Japanese documents they had stolen, so they couldn't judge their importance or how much ransom to demand. Now they can analyze the documents quickly, understand the data's value, and, he speculated, possibly demand higher ransoms as a result. On the language barrier, he said, "that thing is completely gone at this point."

17:57

Malware and Hacking-for-Hire on the Market

Asked whether AI tools are sold on the dark web, Sanchez-Rola first showed a tool called Venom, branded with the Spider-Man villain's face. It is a RAT, a remote access tool, which he explained is not what breaks into a computer but what attackers install afterward to keep ongoing access. Its features include harvesting browser data, turning on the microphone and camera, and running automated tasks.

He then showed a hacking-for-hire service with price tags on a menu of offerings: social media and messenger account takeovers, personal and corporate email hacking, phone hacking, website hacking, university grade changes, personal tracking, "life crashing," and DDoS attacks. These, he said, are professional hackers selling services to anyone who pays.

19:52

AI Makes Cybercrime Easier, Not Newly Possible

The interviewer asked whether AI lets people with no expertise commit cybercrime. Sanchez-Rola reframed it: AI makes cybercrime "easier than ever." People without skills could always find help, and he showed dark web forums where members teach each other techniques. Now they sometimes don't need forums, because AI helps them learn. Mainstream models may refuse requests they recognize as attacks, but unrestricted AI models can be bought or accessed on the dark web and used to carry out attacks.

The forum he showed was selling databases, yet the same board featured a discussion from the previous day about One Piece versus Naruto, sitting alongside threads about stealing and selling people's identities.

21:41

How Ransom Negotiations Happen

Using a non-Japanese case as an example, Sanchez-Rola described what happens after an intrusion. The victim finds a message or file on their computer explaining, in easy step-by-step terms, how to reach the attackers' communication system on the dark web. He showed a chat log: the company asked for an hour, saying it had five days to decide; the attacker replied that it had one hour. The company then offered $100,000 immediately, and the attacker rejected it as not enough, insisting the situation was very bad and they had to pay more. All such negotiations, he said, take place on the dark web.

22:51

The Dark Web Itself Is Not Illegal

Sanchez-Rola said the dark web is not illegal in itself. It was built for private communication, and he stated it was created by US naval research. Besides criminal uses, it has legitimate ones: journalists contacting sources safely, and people in countries where governments block information using it to access that information and organize. Such needs may seem remote in Japan, where information is freely available, but he noted that in many countries this is not the case.

24:13

What One Leaked Password Can Do

Asked what damage a single leaked email and password can cause, Sanchez-Rola said it depends on whether the password is unique and strong. Many people reuse passwords or make small variations, such as adding a dot, a "1," or a capital letter, so one leaked password can open many doors. Some even reuse their bank PIN digits, letting attackers guess bank credentials even when the bank itself wasn't breached. The consequences can include identity theft, bank accounts opened in the victim's name, stolen money, and legal trouble caused by misuse of their identity. Here he recommended dark web monitoring tools "like similar to the ones we offer," adding that receiving an alert is shocking and that having support to calm down and act matters.

On how criminals get from a password to something like a bank account, he said a password and email alone may not be enough, but attackers usually hold much more, such as a passport and address, letting them answer verification questions. They can also impersonate a victim to send a phishing link to friends. Because the message appears to come from someone trusted, people click, so one person's leak can lead to many others being hacked.

27:16

AI-Driven Spear Phishing and Deepfakes

Asked which data is most valuable, Sanchez-Rola said it depends on the goal; everything is useful. My Number, passports, and driver's licenses are obviously important, but logins and addresses matter too. The bigger shift is what AI enables. Spear phishing, highly targeted attacks built from detailed knowledge of a person, used to require someone to dig through the data manually. Now an attacker can feed data on many people into AI, have it build profiles, and generate individualized attacks for each one, making them more efficient and more likely to succeed.

He described how this could stack with voice cloning: an attacker could take audio from something like a voicemail, clone a voice, and call someone using the voice of a person they know, armed with detailed information about both parties. The target may conclude it can't be fake. "It could," he said, which is why people now need to be careful about what they trust.

No Trust Between Criminals, and No Reason to Pay

Asked how buyers and sellers establish trust, Sanchez-Rola answered: "They don't." He advised that victims should not pay. He traced how ransomware evolved: originally attackers encrypted data and demanded payment for access, which worked because people lacked backups. Once people kept backups, attackers shifted to threatening to sell the data. Paying now means trusting the people who hacked you to delete it, with no proof. They might not post it publicly but still sell it privately, collecting money twice. Some long-running groups try to show they are trustworthy by pointing to their longevity, but he said that "means nothing" because there is no way to verify deletion.

32:08

Can Criminals Be Traced?

On whether cryptocurrency payments make attackers untraceable, Sanchez-Rola said tracing is possible. Even without names, investigators can link wallets, and a mistake somewhere in the chain can lead to a person; he said there have been cases caught through Bitcoin payments. Connection errors can also reveal an approximate location, and communication patterns can help. He said Norton supports such efforts and that law enforcement collaborations have taken groups down.

The catch is "mirrors": if the whole infrastructure isn't dismantled, a service may reappear around three weeks later. Seizing servers is less effective than physically catching the people behind them. Tracing is possible but complex, he concluded, since the dark web was built for privacy, and if it were easy these groups wouldn't exist.

33:44

Protecting Individuals and Businesses

Sanchez-Rola said the line between personal and business security is blurry. An employee of a 50-person firm might take a work laptop home, connect remotely over a compromised home network, and create a problem for the company, so businesses must protect employees both inside and outside the office. His recommendations included a password manager with unique passwords, multi-factor authentication, and a "complete" security solution rather than a basic "check the box" product, comparing the latter to blocking your door with paper. He also listed dark web monitoring and identity theft protection.

If something has leaked, the first step depends on what it is. For a password, change it and any password even loosely related to it. For an address, which can't be changed, the value lies in knowing it leaked so you treat unexpected physical letters with more skepticism. For bank-related leaks there are specific processes, and he recommended restoration services, arguing that a shocked person needs a calm guide telling them what to do next, whether a knowledgeable acquaintance or a professional service.

To find out whether data has leaked, he again pointed to dark web monitoring, describing it as a "mini me" scanner that searches forums and sites for your data and reports past and new findings. He reiterated that individuals should not search the dark web themselves: "That's a horrible idea."

For larger companies, he said that lacking a cybersecurity team means "living in the previous century." Small companies often can't afford dedicated experts, so he recommended small-business security services, mentioning that Norton offers them, tailored to that kind of business rather than generic products.

Leaked Data Can't Be Removed

Asked whether leaked data can ever be removed, Sanchez-Rola said essentially no. There is no proof nobody copied it, no way to know how many people obtained it, and no way to know if it has been re-uploaded. He described "scavengers," individuals rather than hacker groups, who collect public leak data from many sources and resell it cheaply, profiting from the effort of gathering it. The realistic response is remediation: changing settings, passwords, and bank arrangements so the leaked data is less useful. Anyone promising deletion, he said, is very unlikely to be able to certify it.

"Reality," Not Paranoia

The interviewer asked whether people should adopt a zero-trust mindset and assume their data has leaked. Sanchez-Rola preferred to call it "reality": with so many breaches, it is unlikely anyone has never interacted with an affected company, though not all of a person's data is necessarily exposed. But he rejected living in paranoia. His rule was to verify whenever money is involved. If a call or message seemingly from your CEO references a real project but something feels off, contact them yourself and ask. "Don't be scared all the time, just be aware."

An Expanding Underground Economy

On where the market is heading, Sanchez-Rola said it is growing, not shrinking. He gave device fingerprinting as an example of techniques seen heavily in recent years: attackers clone a device's fingerprint so that services that flag new locations or new phones believe the login comes from the victim's own machine. As companies add protections, attackers advance to bypass them. He said he can't see the future but considers it very likely that next year's attacks on Japan will be higher than this year's, acknowledging this is not a hopeful message.

43:19

Three Things to Start Today

For individuals, Sanchez-Rola's closing recommendations were concrete. First, use a VPN at all times, noting that phones auto-connect to Wi-Fi networks, such as a café you visited years ago, without you noticing. Second, use a password manager together with two-factor or multi-factor authentication, so a stolen password alone is less useful. Third, use dark web monitoring to understand what of your data is out there. He called a complete security suite so obvious it barely needed mentioning, and added a restoration service as a further recommended step for knowing what to do once data has leaked.