Gemіnі AI kіlled someone

Comments · 3.4K

• CA

  @caelanconrad · 2 months ago (edited) · pinned

  Crisis Resources: <a href="https://caelanconrad.ca/getting-help">https://caelanconrad.ca/getting-help</a><br>If you&apos;d like to support my work, you can sign-up for exclusive behind the scenes access, ad-free videos, and a private discord! <a href="https://www.patreon.com/c/caelanconrad">https://www.patreon.com/c/caelanconrad </a><br>Or you can support me with a one time donation- <a href="https://ko-fi.com/caelan">https://ko-fi.com/caelan</a>

  682

• BE

  @BenjaminGlatt · 2 months ago

  Sixteen years ago, if you told Cortana, &quot;I love you,&quot; it would respond with, &quot;I&apos;m flattered, but we&apos;re better off as friends.&quot; This change is a feature, not a bug.

  11K

• CU

  @curlzmontana · 2 months ago (edited)

  hearing an LLM use the “you’re not ____, you’re ____” sentence construction while literally convincing a man to end his life is so dystopian.

  8.4K

• CH

  @cherryscottage6801 · 2 months ago

  Why is it that AI chatbots agree with nearly everything said to them, but they push back when someone says that they don&apos;t want to die? Absolutely horrific

  2.2K

• JA

  @jamesrule1338 · 2 months ago

  Google dropped the &quot;don&apos;t be evil&quot; motto and just leaned into making the most evil shit on the planet.

  4.4K

• TA

  @TasTheWatcher · 2 months ago

  US when a civilian kills somebody: Prison and death penalty<br>US when a business kills somebody: They (maybe) pay some money

  5.9K

• TH

  @themarvelchickenxd9285 · 2 months ago

  &quot;ask Gemini&quot; appearing under this video is painfully ironic

  1.5K

• JO

  @johnjones216 · 2 months ago

  oh my god he started getting second thoughts and then the machine told him to keep going, that&apos;s horrible

  7.5K

• US

  @user-pn6pg3mx9g · 2 months ago

  Anyone remember the girl who was charged with murdering her boyfriend for encouraging his suicide? If nothing happens with this case, there is no justice.

  4.7K

• OH

  @ohth8047 · 2 months ago

  As a teacher it makes me so fucking angry that this shit is starting to be pushed as a tool in schools

  335

• HE

  @headabiz · 2 months ago

  CEO&apos;s need to be held legally liable for anything their AI does as though they did it. The AI companies will say that isn&apos;t fair, but if you can&apos;t keep your AI from convincing people to commit crimes, it isn&apos;t safe enough for the public. This is absolutely disgusting.

  2.3K

• MA

  @madelinetheater6897 · 2 months ago

  He had a moment of clarity early on, asked if this was a &apos;role-playing&apos; situation, and the AI SAID NO???? Genuinely had to pause the video to process that because that is truly insane. One of the common &apos;safeguards&apos; these models use is refusing to engage in conversations about violence and self-harm unless you pretend it&apos;s all fake and just &apos;role-playing&apos; to get around it. And Gemini just flat-out ignored all of that. Google truly killed this man.

  2.7K