Large Reasoning Models Are Autonomous Jailbreak Agents

Jailbreaking -- bypassing built-in safety mechanisms in AI models -- has traditionally required complex technical procedures or specialized human expertise. In this study, we show that the persuasive capabilities of large reasoning models LRMs simplify and scale jailbreaking, converting it into a...

TelegAI 跨站脚本漏洞

TelegAI is an AI chatbot website from TelegAI, Inc. A cross-site scripting vulnerability exists in TelegAI versions 2025-05-26 and earlier, which stems from an insecure direct object reference that could lead to tampering with other users' conversations and injecting malicious content...

Liner Insecure Direct Object Reference / Brute Force

Liner is a reliable AI search engine with over 10 million users worldwide. It is vulnerable to an insecure direct object reference vulnerability. Conversation histories for all users are stored on the server. However, Liner's server does not distinguish the ownership or sharing status of individu...

Meta AI chatbot bug could have allowed anyone to see private conversations

A researcher has disclosed to TechCrunch that he received a $10,000 bounty for reporting a bug that let anyone access private prompts and responses with the Meta AI chatbot. On June 13, we reported that the Meta AI app publicly exposes user conversations, often without users realizing it. In thes...

ImpReSS: Implicit Recommender System for Support Conversations

Following recent advancements in large language models LLMs, LLM-based chatbots have transformed customer support by automating interactions and providing consistent, scalable service. While LLM-based conversational recommender systems CRSs have attracted attention for their ability to enhance th...

Your Meta AI chats might be public, and it’s not a bug

Conversations that people are having with the Meta AI app are being exposed publicly, often without the users realizing it, revealing a variety of medical, legal, and private matters. The standalone app and the company's integrations with artificial intelligence AI across its platforms—Facebook,...

I

Description Improper authorization controls in the conversation sharing feature make it possible to access other user's conversations given a known conversation ID. The exploitability is limited by the fact that UUIDv4 conversation IDs are generated on the server side and are practically impossib...

OpenAI forced to preserve ChatGPT chats

OpenAI has protested a court order that forces it to retain its users' conversations. The creator of the ChatGPT AI model objected to the order, which is part of a copyright infringement case against it by The New York Times and other publishers. The news organizations argued that ChatGPT was...

CVE-2025-48474

FreeScout is a free self-hosted help desk and shared mailbox. Prior to version 1.8.180, the application incorrectly checks user access rights for conversations. Users with showonlyassignedconversations enabled can assign themselves to an arbitrary conversation from the mailbox to which they have...

CVE-2022-2080

The Sensei LMS WordPress plugin before 4.5.2 does not ensure that the sender of a private message is either the teacher or the original sender, allowing any authenticated user to send messages to arbitrary private conversation via a IDOR attack. Note: Attackers are not able to see...

CVE-2020-9379

The Software Development Kit of the MiContact Center Business with Site Based Security 8.0 through 9.0.1.0 before KB496276 allows an authenticated user to access sensitive information. A successful exploit could allow unauthorized access to user conversations...

CVE-2015-9285

esoTalk 1.0.0g4 has XSS via the PATHINFO to the conversations/ URI...

Malicious code in crisp-import-conversations (npm)

--- -= Per source details. Do not edit below this line.=- Source: ghsa-malware 695e694069b00b1d669921e34a38537e57d625a52cebab80f5ae8af1221b8377 Any computer that has this package installed or running should be considered fully compromised. All secrets and keys stored on that computer should be...

WhatsApp Adds Advanced Chat Privacy to Blocks Chat Exports and Auto-Downloads

WhatsApp has introduced an extra layer of privacy called Advanced Chat Privacy that allows users to block participants from sharing the contents of a conversation in traditional chats and groups. "This new setting available in both chats and groups helps prevent others from taking content outside...

Dust: Privilege Escalation leads to Unauthorized Access to Private Conversations By any Regular user [Read , Edit and Delete]

Summary: A normal authenticated user on dust.tt can escalate their privileges by accessing, modifying, and deleting any chat threads belonging to other users — including administrators — through a vulnerable API endpoint without having the appropriate permissions. Vulnerability Details: Reading...

Unblu Spark 安全漏洞

Unblu Spark is a key component in a conversation-centered digital customer experience platform from Swiss company Unblu. A security vulnerability exists in Unblu Spark that stems from the ability of conversation participants to replace uploaded files...

CVE-2024-13740

The ProfileGrid – User Profiles, Groups and Communities plugin for WordPress is vulnerable to Insecure Direct Object Reference in all versions up to, and including, 5.9.4.2 via the pmmessengershowmessages function due to missing validation on a user controlled key. This makes it possible for...

CVE-2024-13740

The ProfileGrid – User Profiles, Groups and Communities plugin for WordPress is vulnerable to Insecure Direct Object Reference in all versions up to, and including, 5.9.4.2 via the pmmessengershowmessages function due to missing validation on a user controlled key. This makes it possible for...

PT-2025-6605 · WordPress · Profilegrid

Name of the Vulnerable Software and Affected Versions: ProfileGrid – User Profiles, Groups and Communities plugin for WordPress versions up to, and including, 5.9.4.2 Description: The issue allows authenticated attackers with Subscriber-level access and above to read private conversations of othe...

PT-2025-5302 · Apple · Ipados +5

Name of the Vulnerable Software and Affected Versions: visionOS versions prior to 2.3 iOS versions prior to 18.3 iPadOS versions prior to 18.3 macOS Sequoia versions prior to 15.3 watchOS versions prior to 11.3 tvOS versions prior to 18.3 Description: A type confusion issue was addressed with...