The majority of UK security professionals are concerned DeepSeek will open their organizations up to cyber attacks

Thinking of using DeepSeek at work? Think again. Cybersecurity experts have warned you’re putting your enterprise at huge risk.

  • sylver_dragon@lemmy.worldEnglish
    91·
    3 days ago

    As opposed to handing that data over to Google, Microsoft or ChatGPT.
    If anything, DeepSeek almost seems the better choice. You can actually run it disconnected, on your own hardware. Meaning you’re not sending your data off to handled on someone else’s computer. Granted, having not gone over it, it’s possible it contains malware. But, that seems like the type of claim which would need some proof to back up.

    • fitgse@sh.itjust.worksEnglish
      6·
      3 days ago

      The model itself does not contain spyware nor is it filtered. Using the models locally (via ollama or llama.cpp) or in a secure hosted environment like AWS bedrock is safe. I was DeepSeek that way all the time.

      I would never ever recommend using DeepSeek or Alibaba’s Qwen service though!

      • sylver_dragon@lemmy.worldEnglish
        2·
        3 days ago

        I would never ever recommend using DeepSeek or Alibaba’s Qwen service though!

        Ya, fair point. If you put your data on someone else’s computer, it’s no longer your data.

      • tim@infosec.pubEnglish
        1·
        3 days ago

        What about using those services to debug thoroughly shitty personal projects? I have no issue with the Chinese getting their hands on my terrible code as they subsidize my computer science education. It’s crazy how cheap even their reasoner is. Kinda feels like Google/other providers should have their APIs subsidized to the level of Chinas tbh.