More results...
By now, you’ve heard the hype. Agentic AI; self-directed and goal-oriented. Supposedly, the next big thing in security automation. If you’re working in a Security Operations Center (SOC), it might sound like déjà vu. Agentic AI brings autonomous, deci…
Reddit, the popular social media platform known as the “front page of the internet,” has reportedly signed a significant content licensing agreement with an undisclosed AI company. This $60 million deal is poised to impact AI research substantially, Reddit’s impending initial public offering (IPO), and the AI company involved. Document Live Account Takeover Attack Simulation […]
The post Reddit Signed a $60,000,000 Content Licensing Deal with an AI company appeared first on GBHackers on Security | #1 Globally Trusted Cyber Security News Platform.
Multiple critical flaws in the infrastructure supporting AI models have been uncovered by researchers, which raise the risk of server takeover, theft of sensitive information, model poisoning, and unauthorized access. Affected are platforms that are essential for hosting and deploying large language models, including Ray, MLflow, ModelDB, and H20. While some vulnerabilities have been addressed, others have not received a […]
The post Critical AI Tool Vulnerabilities Let Attackers Execute Arbitrary Code appeared first on GBHackers on Security | #1 Globally Trusted Cyber Security News Platform.
Prompt injection refers to a technique where users input specific prompts or instructions to influence the responses generated by a language model like ChatGPT. However, threat actors mainly use this technique to mod the ChatGPT instances for several m…
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 sibli…
A new tool called FraudGPT has been launched by cybercriminals which pose a serious threat to both individuals and businesses. This black-hat-based tool is capable of executing social engineering and Business Email Compromise (BEC) attacks, making it a…
There are rising concerns about the security risks associated with artificial intelligence (AI), which is becoming more and more popular and pervasive. Google, a major participant in the creation of next-generation artificial intelligence (AI), has emp…
The revolutionary innovations by AI (Artificial Intelligence) include generative AI that has various creative potential, but along with that it also raises serious concerns with malicious tools like WormGPT. Since it’s a powerful generative AI-ba…
Cyber criminals have recently started using Facebook to pretend to be well-known generative AI brands like ChatGPT, Google Bard, Midjourney, and Jasper to steal users’ personal information. Users on Facebook are deceived into downloading content …
WormGPT, a black-hat-based tool has been recently launched by cybercriminals and has the potential to conduct various social engineering as well as Business Email Compromise (BEC) attacks. This tool has no limitations towards its use and has no boundar…