X-Industry

Autonomous vehicles and many other automated systems are controlled by AI, but the AI itself could be compromised by malicious attackers who take control of the AI’s weights.  Weights within AI’s deep neural networks represent the models’ learning and how it is used.  A weight is usually defined in a 32-bit word, and there can be hundreds of billions of bits involved in this AI's reasoning process.  It is a no-brainer that if an attacker controls the weights, they control the AI.[1]

A research t

The cybersecurity company ESET has disclosed that it discovered an artificial intelligence (AI)-powered ransomware variant codenamed PromptLock.  Written in Golang, the newly identified strain uses the gpt-oss:20b model from OpenAI locally via the Ollama API to generate malicious Lua scripts in real-time.  The open-weight language model was released by OpenAI earlier this month.  "PromptLock leverages Lua scripts generated from hard-coded prompts to enumerate the local filesystem, inspect target

A recent report by Salt Security highlights a critical warning: without proper Application Programming Interface (API) discovery, governance, and security, the very technology meant to drive smarter customer engagement could open the door to cyber-attacks or data leakage.  The research also reveals an increasing trust gap between businesses that deploy agentic AI for external communications and consumers who are wary of sharing personal information due to security concerns.

Because APIs power AI

A proof-of-concept attack detailed by Neural Trust demonstrates how bad actors can manipulate LLMs into producing prohibited content without issuing an explicitly harmful request.  Named "Echo Chamber," the exploit uses a chain of subtle prompts to bypass existing safety guardrails by manipulating the model's emotional tone and contextual assumptions.  Developed by Neural Trust researcher Ahmad Alobaid, the attack hinges on context poisoning.  Rather than directly asking the model to generate in

A proof-of-concept attack detailed by Neural Trust demonstrates how bad actors can manipulate LLMs into producing prohibited content without issuing an explicitly harmful request.  Named "Echo Chamber," the exploit uses a chain of subtle prompts to bypass existing safety guardrails by manipulating the model's emotional tone and contextual assumptions.  Developed by Neural Trust researcher Ahmad Alobaid, the attack hinges on context poisoning.  Rather than directly asking the model to generate in

A proof-of-concept attack detailed by Neural Trust demonstrates how bad actors can manipulate LLMs into producing prohibited content without issuing an explicitly harmful request.  Named "Echo Chamber," the exploit uses a chain of subtle prompts to bypass existing safety guardrails by manipulating the model's emotional tone and contextual assumptions.  Developed by Neural Trust researcher Ahmad Alobaid, the attack hinges on context poisoning.  Rather than directly asking the model to generate in

A proof-of-concept attack detailed by Neural Trust demonstrates how bad actors can manipulate LLMs into producing prohibited content without issuing an explicitly harmful request.  Named "Echo Chamber," the exploit uses a chain of subtle prompts to bypass existing safety guardrails by manipulating the model's emotional tone and contextual assumptions.  Developed by Neural Trust researcher Ahmad Alobaid, the attack hinges on context poisoning.  Rather than directly asking the model to generate in

A proof-of-concept attack detailed by Neural Trust demonstrates how bad actors can manipulate LLMs into producing prohibited content without issuing an explicitly harmful request.  Named "Echo Chamber," the exploit uses a chain of subtle prompts to bypass existing safety guardrails by manipulating the model's emotional tone and contextual assumptions.  Developed by Neural Trust researcher Ahmad Alobaid, the attack hinges on context poisoning.  Rather than directly asking the model to generate in

A proof-of-concept attack detailed by Neural Trust demonstrates how bad actors can manipulate LLMs into producing prohibited content without issuing an explicitly harmful request.  Named "Echo Chamber," the exploit uses a chain of subtle prompts to bypass existing safety guardrails by manipulating the model's emotional tone and contextual assumptions.  Developed by Neural Trust researcher Ahmad Alobaid, the attack hinges on context poisoning.  Rather than directly asking the model to generate in

A proof-of-concept attack detailed by Neural Trust demonstrates how bad actors can manipulate LLMs into producing prohibited content without issuing an explicitly harmful request.  Named "Echo Chamber," the exploit uses a chain of subtle prompts to bypass existing safety guardrails by manipulating the model's emotional tone and contextual assumptions.  Developed by Neural Trust researcher Ahmad Alobaid, the attack hinges on context poisoning.  Rather than directly asking the model to generate in

A proof-of-concept attack detailed by Neural Trust demonstrates how bad actors can manipulate LLMs into producing prohibited content without issuing an explicitly harmful request.  Named "Echo Chamber," the exploit uses a chain of subtle prompts to bypass existing safety guardrails by manipulating the model's emotional tone and contextual assumptions.  Developed by Neural Trust researcher Ahmad Alobaid, the attack hinges on context poisoning.  Rather than directly asking the model to generate in

A proof-of-concept attack detailed by Neural Trust demonstrates how bad actors can manipulate LLMs into producing prohibited content without issuing an explicitly harmful request.  Named "Echo Chamber," the exploit uses a chain of subtle prompts to bypass existing safety guardrails by manipulating the model's emotional tone and contextual assumptions.  Developed by Neural Trust researcher Ahmad Alobaid, the attack hinges on context poisoning.  Rather than directly asking the model to generate in