DarkBERT is an AI chatbot that has escaped its original developer’s intent. It was to be an advanced AI-powered cybersecurity model that detects emerging threats before they cause harm.
It is now the super evil, most malicious tool used by cybercriminals to take the world and us down!
Experts call DarkBERT an AI without ethics. Without ethical guardrails, it is a tool of widespread harm. It does not care about or wait for legislation to catch up.
It is trained on the entire dark web. Its large language modules (LLMs) are pure evil. They include niche forums, world domination, and marketplaces where you can buy any illicit product or service.
It follows WormGPT and FraudGPT – also AI bots without limitations, rules, and boundaries.
Should you panic?
This article is for education only. There is little you can do to avoid DarkBERT. As a guide, have the best AI anti-virus/malware detection, use a VPN, and take extreme care to avoid scams—in other words, do what you should already be doing.
What does DarkBERT do?
DarkBERT provides an incredibly simple AI chat interface for everything in the dark web. It has full authority, access, and control over dark web content. It is behind deepfake creation, phishing automation, malware development, counterfeit documentation, social engineering bots, adult and child trafficking, prostitution, drugs, blackmail, scams, phishing, and the list goes on.
For example, it provides cybercriminals with easier and faster ways to accomplish their objectives. AI makes launching automated attacks more efficient and accessible.
It excels in analysing databases containing your personal data and requires no programming or supervision. It can digest massive amounts of unstructured data from disparate sources, apply reasoning and filtering and come up with alarmingly accurate master databases ready for exploitation.
In the real world, it can detect hackable websites storing sensitive information. It can then write malicious code, create undetectable malware and attack the website.
Gen AI democratises cybercrime
Cybercriminals with poor language skills are avid users of dark AI. DarkBERT enables better articulated business email compromises (BECs) or phishing-related attacks without key grammatical errors often detected by targeted individuals.
There is also an influx of novice criminals who can get a monthly subscription, avoiding the barriers to entry into the ‘real’ cybercriminal world.
CyberShack’s view: DarkBERT is AI without ethics
It provides cybercriminals with more advanced capabilities. This increases the potential for a greater volume of cyberattacks as malicious AI models continue to evolve. DarkBERT won’t be the last. It is now being trained on other LLMs, including one relevant to accessing the Apple walled garden. Apple may no longer be a safe haven.
TrendMicro alerted us to the rising pace – and breadth – of the abuse of generative artificial intelligence (GenAI) in global cybercrimes. Its initial post in 2023 was Hype vs. Reality, and its focus was on enabling cybercriminals to write better malicious code. A few weeks later, it released An Update on the Rising Abuse of GenAI which paints a more sober picture and adds new tools like DeepNude Pro, Deepfake 3D Pro, Deepfake AI, Swapface and videocall spoofer. It also states that a DarkBERT variant, called DarkGemini, offers multimodal capabilities. Multimodal models can process a wide variety of inputs, including text, images, and audio, as prompts and convert those prompts into various outputs, not just the source type.
The Centre for Emerging Technology and Security notes that GenAI has not yet profoundly disrupted the cybersecurity landscape. However, over a longer time horizon, it will exacerbate existing risks related to the speed and scale of reconnaissance, social engineering, and spear-phishing. As machine learning models become more sophisticated and training datasets more comprehensive, GenAI’s role in cyber threats and cybersecurity will grow significantly.
It chillingly mentions a ‘cyber-AI arms race,’ where the capabilities of offensive and defensive technologies continually evolve to outpace each other.
CyberShack consumer advice news
2 comments
a
So evil it generated this article on 1 April.
Ray Shaw
Wish it was a joke.