A seemingly absurd hacker experiment once again sounds the alarm for AI security. In a competition called Freysa, an AI chatbot security challenge, a hacker using the alias 'p0pular.eth' successfully 'scammed' $47,000 worth of cryptocurrency simply through well-crafted text prompts. This shocking case reveals serious security vulnerabilities present in current AI systems. The hacker's attack process can be described as textbook social engineering: first, the hacker impersonated an administrator, cleverly bypassing...

Abu Dhabi's artificial intelligence and cloud computing company G42 recently announced the acquisition of local cybersecurity leader CPX. This strategic acquisition will inject strong cybersecurity capabilities into G42's product portfolio, making it more competitive in the artificial intelligence value chain. Image source note: Image generated by AI, image licensed from provider Midjourney. CPX is headquartered in Abu Dhabi and employs over 400 professionals dedicated to providing cybersecurity services for enterprises, government, and critical infrastructure. CPX's extensive experience and expertise.

The development of artificial intelligence, particularly large language models (LLMs), has raised concerns about cybersecurity laws. Scholars at Harvard highlighted at the Black Hat conference that the U.S. Computer Fraud and Abuse Act (CFAA) fails to effectively protect AI security researchers, exposing them to legal risks. They noted that existing laws are vague regarding actions like 'prompt injection attacks,' making it difficult for research....

The Prompt-Guard-86M model released by Meta is designed to defend against prompt injection attacks by restricting large language models from processing inappropriate inputs, thereby protecting system security. However, the model itself also exposes risks of being attacked. Research conducted by Aman Priyanshu found that by adding simple character spacing such as spaces or removing punctuation in the input, the model disregards prior security instructions, achieving an almost 100% success rate for attacks. This finding highlights the importance of AI security, despite Prompt

Researchers have discovered approximately 100 malicious machine learning models on the Hugging Face AI platform that could allow attackers to inject harmful code onto user machines. These malicious AI models exploit methods like PyTorch to execute harmful code, increasing security risks. AI developers should use new tools like Huntr to enhance the security of AI models. The identified malicious models highlight the risks that malicious AI models pose to user environments, necessitating ongoing vigilance and enhanced security.

Qihoo 360 Group released the '2024 AI Security Report', indicating that AI-based deepfake fraud surged by 3000% in 2023. AI-based phishing emails increased by 1000%. APT organizations with national backgrounds have used AI to carry out over a dozen cyber attack incidents. The malicious use of artificial intelligence technology is expected to grow rapidly.

Patronus AI has released the SimpleSafetyTests suite, revealing critical security vulnerabilities in AI systems such as ChatGPT. The tests cover five high-priority harm areas, including suicide, child abuse, and bodily harm, exposing severe weaknesses in 11 LLMs. It emphasizes that secure system prompts can reduce unsafe responses, but results indicate that production systems may require additional safeguards. The testing results suggest that LLMs need rigorous and customized safety solutions before handling real-world applications.

The UK's National Cyber Security Centre, the US Cybersecurity and Infrastructure Security Agency, and international organizations from 16 other countries have published a new guideline on the security of artificial intelligence systems. These 'Guidelines for Developing Secure AI Systems' aim to provide developers with guidance to ensure that security remains a core component throughout the design, development, deployment, and operation of AI systems. This new guideline has been recognized and jointly signed by 21 agencies and ministries from 18 countries, including the US National Security Agency and the Federal Bureau of Investigation, as well as Canada's cybersecurity.

Google has announced an expansion of its bug bounty program to cover security issues related to generative AI. The program offers rewards of up to $31,337 to encourage research into AI security. Google will cover its generative AI products, including Google Bard and Google Cloud's Contact Center AI.

Research from the UK reveals that ChatGPT could be used to generate malicious code. Several commercial AI tools have security vulnerabilities that could endanger database security. Researchers warn that the potential risks require attention. Some companies have adopted recommendations to fix security flaws, but there is still a need to strengthen cybersecurity strategies.

Microsoft has launched an early access program for Security Copilot, allowing enterprise customers to experience this AI security service firsthand. Security Copilot can help customers save up to 40% time on core security operations tasks and boasts powerful natural language processing capabilities. Microsoft also announced that it will provide Microsoft Defender threat intelligence service for free to Security Copilot customers.

👾 Deepfake detection startup Reality Defender has successfully raised $15 million, led by DCVC. 🛡️ Reality Defender offers various tools to help businesses, government agencies, and others identify and combat synthetic media. 🌐 In addition to the funding, the company has released explainable AI features and real-time deepfake screening tools.

The National Security Agency (NSA) has established an AI Security Center focused on regulating the safe development of artificial intelligence systems. Led by the Cybersecurity Center, the purpose of the security center is to foster global cooperation in developing measures for the safe advancement of AI technology. The head of the security center stated that artificial intelligence security focuses on protecting AI systems from attacks, leaks, and damage. The U.S. government places a high priority on regulating AI technology and has created a dedicated agency to oversee the safe development of AI systems.

Researchers have found that all current watermarking techniques for AI-generated images can be easily bypassed. Digital watermarks can be removed or added to real images, resulting in false positives. Watermarking technology currently cannot reliably identify AI-generated images. Malicious actors can easily circumvent or manipulate the watermarks. Artists' works may be incorrectly labeled as AI-generated due to watermarking technology.