Zeitpunkt Nutzer Delta Tröts TNR Titel Version maxTL Sa 28.09.2024 00:00:44 7.208 0 641.886 89,1 NerdCulture 4.2.12 1.000 Sa 28.09.2024 00:00:44 7.208 0 641.886 89,1 NerdCulture 4.2.12 1.000 Fr 27.09.2024 00:01:09 7.208 -1 641.238 89,0 NerdCulture 4.2.12 1.000 Do 26.09.2024 00:00:08 7.209 -1 640.599 88,9 NerdCulture 4.2.12 1.000 Mi 25.09.2024 00:00:05 7.210 -3 640.317 88,8 NerdCulture 4.2.12 1.000 Di 24.09.2024 00:01:09 7.213 0 639.780 88,7 NerdCulture 4.2.12 1.000 Mo 23.09.2024 00:00:50 7.213 0 639.042 88,6 NerdCulture 4.2.12 1.000 So 22.09.2024 00:00:07 7.213 -2 638.417 88,5 NerdCulture 4.2.12 1.000 Sa 21.09.2024 00:01:12 7.215 -1 637.765 88,4 NerdCulture 4.2.12 1.000 Fr 20.09.2024 00:01:14 7.216 0 637.248 88,3 NerdCulture 4.2.12 1.000
Olly 👾 (@Olly42) · 01/2024 · Tröts: 312 · Folger: 35
Sa 28.09.2024 14:33
[ChatGPT macOS Flaw] Long-Term Spyware via Memory Function. 🕵️♂️ :apple_inc:
A now-patched security vulnerability in OpenAI's ChatGPT app for macOS could have made it possible for attackers to plant long-term persistent spyware into the artificial intelligence (AI) tool's memory.
https://embracethered.com/blog/posts/2024/chatgpt-macos-app-persistent-data-exfiltration/
#apple #macos #openai #chatgpt #llm #spyware #flaw #artificialintelligence #it #security #privacy #engineering #tech #media #news
The technique, dubbed SpAIware, could be abused to facilitate "continuous data exfiltration of any information the user typed or responses received by ChatGPT, including any future chat sessions," security researcher Johann Rehberger said. The issue, at its core, abuses a feature called memory, which OpenAI introduced earlier this February before rolling it out to ChatGPT Free, Plus, Team and Enterprise users at the start of the month. What it does is essentially allow ChatGPT to remember certain things across chats so that it saves users the effort of repeating the same information over and over again. Users also have the option to instruct the program to forget something.
[ImageSource: embracethered.com] The attack technique also builds on prior findings that involve using indirect prompt injection to manipulate memories so as to remember false information, or even malicious instructions, achieving a form of persistence that survives between conversations. "ChatGPT's memories evolve with your interactions and aren't linked to specific conversations," OpenAI says. "Deleting a chat doesn't erase its memories; you must delete the memory itself."
[Öffentlich] Antw.: 0 Wtrl.: 0 Fav.: 0 · via Metatext