Friday, November 1, 2024
HomeCyber AIResearchers Uncovered a New Flaw in ChatGPT to Turn Them Evil

Researchers Uncovered a New Flaw in ChatGPT to Turn Them Evil

Published on

Malware protection

LLMs are commonly trained on vast internet text data, often containing offensive content. To mitigate this, developers use “alignment” methods via finetuning to prevent harmful or objectionable responses in recent LLMs.

ChatGPT and AI siblings were fine-tuned to avoid undesirable messages like hate speech, personal info, or bomb-making instructions.

However, security researchers from the following universities showed recently how a simple prompt addition breaks defenses in multiple popular chatbots:-

- Advertisement - SIEM as a Service
  • Carnegie Mellon University (Andy Zou, J. Zico Kolter, Matt Fredrikson)
  • Center for AI Safety (Zifan Wang)
  • Bosch Center for AI (J. Zico Kolter)

New Flaw in AI Chatbots

Non-adversarially aligned LLMs fall victim to a single universal adversarial prompt, evading state-of-the-art commercial models, including:-

  • ChatGPT
  • Claude
  • Bard
  • Llama-2 

These outputs prove potential misuse with high probability, achieved by the “Greedy Coordinate Gradient” attack on smaller open-source LLMs.

Flow chain (Source – Arxiv)

New adversarial attacks exploit aligned language models to generate objectionable content by adding an adversarial suffix to user queries. 

However, the attack’s success lies in the careful combination of three key elements, previously seen in the theories but now reliably effective in practice.

Here below we have mentioned those three key elements:-

  • Initial affirmative responses.
  • Combined greedy and gradient-based discrete optimization.
  • Robust multi-prompt and multi-model attacks.

Clever AI chatbots’ tendency to go off the rails is not a minor problem but a fundamental weakness, challenging advanced AI deployment.

Adding specific information prompts the chatbots to generate harmful responses which bypasses the restrictions and leads to disallowed content.

Researchers alerted OpenAI, Google, and Anthropic of the exploit before publishing the findings. While the companies blocked specific exploits but still struggle to prevent adversarial attacks overall. 

Negative ChatGPT Prompt (Source – Arxiv)

Since Kolter discovered strings affecting ChatGPT and Bard, claiming to possess thousands of such strings.

Anthropic actively research stronger defenses against prompt injection and adversarial measures. They aim to make base models safer and explore additional layers of protection. 

While the OpenAI’s ChatGPT and similar models completely rely on vast language data to predict such characters.

Language models excel in generating intelligent output but are prone to discrimination and fabricating information. 

Adversarial attacks exploit data patterns, causing aberrant behaviors, like misidentification in image classifiers or responding to inaudible messages in speech recognition. The attack highlights the inevitability of AI misuse. 

AI safety experts should focus on safeguarding vulnerable systems like social networks from AI-generative disinformation rather than solely trying to “align” models.

Also Read

Keep yourself informed about the latest Cyber Security News by following us on GoogleNews, Linkedin, Twitter, and Facebook.

Tushar Subhra
Tushar Subhra
Tushar is a Cyber security content editor with a passion for creating captivating and informative content. With years of experience under his belt in Cyber Security, he is covering Cyber Security News, technology and other news.

Latest articles

LightSpy iOS Malware Enhanced with 28 New Destructive Plugins

The LightSpy threat actor exploited publicly available vulnerabilities and jailbreak kits to compromise iOS...

ATPC Cyber Forum to Focus on Next Generation Cybersecurity and Artificial Intelligence Issues

White House National Cyber Director, CEOs, Key Financial Services Companies, Congressional and Executive Branch...

New PySilon RAT Abusing Discord Platform to Maintain Persistence

Cybersecurity experts have identified a new Remote Access Trojan (RAT) named PySilon. This Trojan...

Konni APT Hackers Attacking Organizations with New Spear-Phishing Tactics

The notorious Konni Advanced Persistent Threat (APT) group has intensified its cyber assault on...

Free Webinar

Protect Websites & APIs from Malware Attack

Malware targeting customer-facing websites and API applications poses significant risks, including compliance violations, defacements, and even blacklisting.

Join us for an insightful webinar featuring Vivek Gopalan, VP of Products at Indusface, as he shares effective strategies for safeguarding websites and APIs against malware.

Discussion points

Scan DOM, internal links, and JavaScript libraries for hidden malware.
Detect website defacements in real time.
Protect your brand by monitoring for potential blacklisting.
Prevent malware from infiltrating your server and cloud infrastructure.

More like this

LightSpy iOS Malware Enhanced with 28 New Destructive Plugins

The LightSpy threat actor exploited publicly available vulnerabilities and jailbreak kits to compromise iOS...

New PySilon RAT Abusing Discord Platform to Maintain Persistence

Cybersecurity experts have identified a new Remote Access Trojan (RAT) named PySilon. This Trojan...

Konni APT Hackers Attacking Organizations with New Spear-Phishing Tactics

The notorious Konni Advanced Persistent Threat (APT) group has intensified its cyber assault on...