What Is DIG AI? Dark Web’s Criminal AI, Explained
A new kind of artificial intelligence has appeared on the dark web, and security researchers are worried about it. It is called DIG AI, and unlike the assistants most people use, it was built with no rules. This guide explains what DIG AI is, what researchers say it can do, whether it is legal, and — most importantly — how ordinary people can stay safe from the crimes it helps power.
One clarification first. This article is about the criminal dark-web tool. It is not dig.ai, the legitimate social-analytics company, and it is not DIG-AI, an academic plant-research project. Those simply share a name. At aiera.blog we cover AI safety and consumer protection, so our focus is the threat and your defence — not access. We do not link to the tool and never will.
What is DIG AI?
DIG AI is an uncensored AI chatbot that operates on the dark web and is marketed to criminals. It first drew public attention when the threat-intelligence firm Resecurity identified it in September 2025 and reported a sharp rise in its use through the final quarter of the year. Its operator, who uses the alias “Pitch,” claims the tool is built on a commercial chatbot engine with its safety limits stripped out.
The case has since been logged as an AI incident by OECD.AI, the international policy body that tracks real-world AI harms. In short: this is not a rumour or a hoax. It is a documented tool that reputable researchers have tested and warned about.
DIG AI vs. the AI you already know
| Name | What it is |
|---|---|
| DIG AI | An uncensored dark-web tool marketed for crime (the subject of this article) |
| dig.ai | A legitimate, unrelated social and brand-intelligence company |
| DIG-AI | An unrelated academic plant-genomics research project |
How is DIG AI different from ChatGPT, Claude, or Gemini?
Mainstream assistants such as ChatGPT, Claude, and Gemini run content-moderation systems that refuse dangerous requests — writing malware, planning fraud, producing abuse material. Those refusals exist because of law, ethics, and regulation.
DIG AI belongs to a category researchers call “criminal AI”: models designed specifically to bypass those guardrails. The difference is not the underlying technology — it is the deliberate removal of every safety layer, so the tool answers requests a legitimate assistant would reject.
What researchers say DIG AI can do
Based on Resecurity’s testing, DIG AI can generate content across a range of illegal areas, including fraud schemes, malicious code, and scam automation. Researchers reported that it requires no registration and that advertising for it has appeared on underground marketplaces already linked to drug trafficking and stolen payment data.
The concern is not that it invents new crimes. It is that it lowers the skill barrier. Tasks that once needed a technical criminal can now be attempted by someone with none — faster, cheaper, and at greater scale. That is the real story behind DIG AI: automation of harm.
The most serious concern: AI-generated abuse material
The gravest finding involves child sexual abuse material (CSAM). Resecurity reported that DIG AI can assist in producing or manipulating illegal content involving minors. Law-enforcement agencies worldwide are already reporting a rise in AI-generated CSAM, including synthetic images used for extortion.
Several jurisdictions — including the EU, the UK, and Australia — now criminalise AI-generated CSAM regardless of whether a real child is depicted. If you ever encounter such material, do not investigate it yourself. Report it to a dedicated body such as the Internet Watch Foundation, which exists to handle exactly these reports safely.
Why the dark web breaks AI safety rules
Mainstream providers block harmful outputs partly because rules such as the EU AI Act require it. Those safeguards, however, do not reach services hosted anonymously on the dark web, which sit outside normal legal and jurisdictional boundaries.
The wider trend is that criminals are taking open-weight AI models, stripping out the safety filters, and re-training them to produce illegal outputs on demand. DIG AI is not the first of its kind — it follows earlier criminal tools such as WormGPT and FraudGPT — and it represents an emerging underground economy of “AI-as-a-service” for crime.
Is DIG AI legal?
Using a tool like DIG AI to create malware, run scams, or generate abuse material is a crime in most countries, and AI-generated CSAM is illegal in a growing number of jurisdictions regardless of how it was made. The tool’s anonymity does not change the law; it only makes enforcement harder.
This section is general information, not legal advice. If you have a specific legal question, consult a qualified professional in your country.
How DIG AI affects ordinary people — and how to protect yourself
You do not need to touch the dark web to be affected by it. The outputs of criminal AI reach ordinary inboxes and phones as sharper phishing emails, convincing scam messages, and realistic fakes. A few habits close most of that gap:
- Slow down on urgency. Any message pushing you to send money, share a code, or “act now” deserves suspicion. Verify through a second, known channel before acting.
- Turn on multi-factor authentication (MFA) on email, banking, and social accounts. It blocks most account takeovers even if a password leaks.
- Learn to spot synthetic media. AI text, images, and audio are getting harder to identify — our guide on whether AI detectors are accurate explains what these tools can and cannot catch.
- Watch for voice-cloning scams. A cloned voice of a relative asking for money is now realistic and cheap to make; our explainer on AI voice cloning shows how the technology works. Agree a family “safe word” for emergencies.
- Talk to your family, especially children, about synthetic images and sextortion, and keep the reporting routes above handy.
The bigger picture
DIG AI is best understood as an early warning, not an isolated event. As criminal and extremist actors adopt tools like it, defenders face threats that operate at greater speed and scale than before. The response is not panic — it is awareness plus basic security hygiene, applied consistently.
Frequently asked questions
Is DIG AI real?
Yes. It was identified and tested by the threat-intelligence firm Resecurity and logged as an AI incident by OECD.AI. It is a documented dark-web tool, not a rumour.
Is DIG AI the same as dig.ai?
No. dig.ai is a legitimate, unrelated social-analytics company. The criminal tool and the legitimate company only share a name.
Is DIG AI legal to use?
Using it to create malware, fraud, or abuse material is a crime in most countries, and AI-generated CSAM is illegal in a growing number of jurisdictions regardless of how it is produced.
How is DIG AI different from ChatGPT?
Tools like ChatGPT refuse dangerous requests through built-in safety systems. DIG AI is designed to remove those safeguards so it will answer requests a legitimate assistant rejects.
Can DIG AI be shut down?
It is difficult. Dark-web services are hosted anonymously and sit outside normal jurisdictions, which is exactly why researchers stress protection and awareness over waiting for a takedown.
The bottom line
DIG AI shows how quickly artificial intelligence can be turned against people once its safety limits are removed. You cannot control the tool, but you can control your exposure to what it produces. Stay sceptical of urgency, switch on MFA, learn the signs of synthetic media, and you neutralise most of the risk it creates.
