home.social

#inboxes — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #inboxes, aggregated by home.social.

  1. New #attack on #ChatGPT research agent pilfers secrets from #Gmail #inboxes

    The face-palm-worthy #prompt injections against #AI assistants continue. Today’s installment hits #OpenAI ’s #DeepResearch agent. Researchers recently devised an attack that plucked #confidential information out of a user’s Gmail inbox and sent it to an attacker-controlled web server, with no interaction required on the part of the victim and no sign of exfiltration.
    #privacy #promptinjection

    arstechnica.com/information-te

  2. New #attack on #ChatGPT research agent pilfers secrets from #Gmail #inboxes

    The face-palm-worthy #prompt injections against #AI assistants continue. Today’s installment hits #OpenAI ’s #DeepResearch agent. Researchers recently devised an attack that plucked #confidential information out of a user’s Gmail inbox and sent it to an attacker-controlled web server, with no interaction required on the part of the victim and no sign of exfiltration.
    #privacy #promptinjection

    arstechnica.com/information-te

  3. New #attack on #ChatGPT research agent pilfers secrets from #Gmail #inboxes

    The face-palm-worthy #prompt injections against #AI assistants continue. Today’s installment hits #OpenAI ’s #DeepResearch agent. Researchers recently devised an attack that plucked #confidential information out of a user’s Gmail inbox and sent it to an attacker-controlled web server, with no interaction required on the part of the victim and no sign of exfiltration.
    #privacy #promptinjection

    arstechnica.com/information-te

  4. New on research agent pilfers secrets from

    The face-palm-worthy injections against assistants continue. Today’s installment hits ’s agent. Researchers recently devised an attack that plucked information out of a user’s Gmail inbox and sent it to an attacker-controlled web server, with no interaction required on the part of the victim and no sign of exfiltration.

    arstechnica.com/information-te

  5. New #attack on #ChatGPT research agent pilfers secrets from #Gmail #inboxes

    The face-palm-worthy #prompt injections against #AI assistants continue. Today’s installment hits #OpenAI ’s #DeepResearch agent. Researchers recently devised an attack that plucked #confidential information out of a user’s Gmail inbox and sent it to an attacker-controlled web server, with no interaction required on the part of the victim and no sign of exfiltration.
    #privacy #promptinjection

    arstechnica.com/information-te

  6. There will be #GoodScotch in #TheFuture...

    #IT's a #Prophecy... #CaptainRedBeanBearSaysSo...

    There will also #BeThings... #InBoxes...

    And, #MoreThings in #Boxes... #InBoxes... #InBoxes...

    🧙:PirateBadge:🤖:wolfparty:🤖:PirateBadge:🧙 | :fediverse:​🦹🔮🐻🔮🦹:fediverse: