WormGPT: The Dark Side of AI Rivaling ChatGPT

In the artificial intelligence space, the rise of generative AI tools such as OpenAI’s ChatGPT has brought bitter-sweet reactions. Recently, though, a darker side of this technology has made its appearance in that WormGPT has launched as a competitor to ChatGPT. Much like its competitor, WormGPT is also a generative AI tool. Instead of aiding dialogue generation, it has been earmarked by cybercriminals for malicious reasons.

SlashNext, founded by Atif Mushtaq in 2015, is a cybersecurity firm that leverages artificial intelligence and behavioural analysis to detect and counteract sophisticated multi-channel phishing attacks. During a significant rise in phishing, morphing payloads, and a very strong belief that phishing would grow from email into a multi-channel problem and become the number one cyber threat to organisations. There was a recognition that something needed to change.

Atif Mushtaq, the visionary founder behind SlashNext and one of the main architects of FireEye’s core malware sandbox technology, recognized phishing detection, like malware, needed to move from signature-based reputation techniques to behavioral analysis. What was needed was a faster, more automated, and accurate approach to phishing detection. With that knowledge, SlashNext built an Artificial Intelligence (AI) phishing detection engine with virtual browsers and behavioral analysis using computer vision, natural language processing, and several machine-learning classifiers to detect and stop well-crafted multi-payload phishing attacks delivered multiple communication channels, including legitimate infrastructure.

Today, as humans have moved completely to the cloud, using apps, browsers and mobile devices to communicate with work, family, and friends, SlashNext is uniquely positioned to stop multi-channel phishing and human hacking.

Unlike ChatGPT, WormGPT is allegedly trained on a wide range of data, with a particular emphasis on malware-related information. This specific training makes it an efficient weapon in the armoury of cybercriminals, equipping them with the ability to generate personalised, deceptive emails.

WormGPT inherits some remarkable features from its base model, GPT-J. These features include the capability to support unlimited characters, retain chat memory, and offer code formatting capabilities. Collectively, these features make the tool a potent ally for those intending to execute sophisticated phishing or business email compromise (BEC) attacks.


The Real-Life Implications


The existence and deployment of WormGPT pose a significant threat to the cyber world. Tests conducted by SlashNext revealed how effectively this AI model can be utilised to engineer convincing phishing attempts. For example, the AI was able to generate an email persuading an unsuspecting account manager into paying a fraudulent invoice, demonstrating the potential of such tools when utilised maliciously.

The reality of AI-generated attacks also highlights an alarming development: the democratisation of advanced BEC attacks. With tools like WormGPT readily available, even cybercriminals with limited skills can carry out sophisticated attacks. More so, these AI tools can generate grammatically flawless emails, reducing the chances of them being flagged as suspicious, and thereby increasing the likelihood of successful attacks.


WormGPT vs ChatGPT


WormGPT as a ‘dark twin’ to ChatGPT is indicative of how AI technology can be adapted for illicit activities. While ChatGPT and similar tools come with safeguards to prevent them from encouraging or engaging in harmful actions, WormGPT has been designed explicitly for these activities.

The launch of WormGPT brings into focus the reality that alongside technological progress comes the risk of such tools being misappropriated. This once again highlights the need for vigilance, well-informed cybersecurity strategies, and an ongoing commitment to adapt in the face of new challenges. Staying informed and alert in this everchanging digital realm is the best solution towards creating safe digital spaces for ourselves and our businesses, as entrepreneurs and professionals in this realm.