lorry (@lorry@infosec.exchange)
I wonder how much damage ChatGPT (in this case, 5.2) causes to people who follow its instructions without checking.
In this case, I saw where it was heading and was curious if it would just tell me to wipe out all my files, so I went with it. And it did.
It's told me to do some pretty destructive things in the past which I have caught, but casually wiping out an entire instance with all of its user-data and trying to frame that as "good news" and pointing out that I should "Accept the loss as done" (which to be fair, would be the only possible action in this case) is about as dangerous as it gets.
#ChatGPT #Dangerous #AI #Safety #Tech
Show Original Post
roy_baer (@roy_baer@norden.social)
@ginnie @HeidrunMueller Zu Anfang Adé zu #Apple / #IOS , #Google , #ChatGPT , #Amazon , #Microsoft , #Facebook , #Instagram , #WhatsApp , #Intel , #AMD , #Seagate , #WesternDigital , #HP , #Azure , #McrosoftOffice / #Office365 , #Cisco , #Fortinet , #Zscaler , #Checkpoint, #PaloAlto , #Adobe - dann natürlich weg von fossilen Brennstoffen, weg vom Dollar & atomare Aufrüstung der EU wird wohl alternativlos sein.
Show Original Post
LucCogZest (@LucCogZest@fediscience.org)
Today, using Script Debugger and #ChatGPT, I wrote and debugged an #AppleScript for use as part of a #Hazel rule. I hooked the script to the AI chat for future reference. Another huge use case for #hookmark.
Unfortunately, Hazel is not [link-friendly](https://hookproductivity.com/help/integration/data-linkability-and-why-it-matters/) so I couldn't hook these two resources to the Hazel rule.
Show Original Post
TechDesk (@TechDesk@flipboard.social)
From @WIRED: The U.S. invaded Venezuela and captured Nicolás Maduro. ChatGPT’s output delivered something else. Read more on how some AI chatbots have a surprisingly good handle on breaking news, while others don’t:
#Tech #ChatGPT #AI #Technology #ArtificialIntelligence
Show Original Post
emmecola (@emmecola@mastodon.uno)
Ho chiesto all'AI se le azioni di #Trump in #Venezuela siano legali. Ecco le risposte.
#ChatGPT: "Trump’s actions sit on shaky legal ground under both US law and international law"
#Claude: "Trump's military action appears to violate both US and international law"
#Grok: "Given what is publicly known today, “legal” is not something that can be stated with certainty by a neutral observer—but the strongest reading of the standard legal frameworks points toward “probably not legal""
Show Original Post
PrivacyDigest (@PrivacyDigest@mas.to)
The US Invaded #Venezuela and Captured #NicolásMaduro. #ChatGPT Disagrees
Some #AI #chatbots have a surprisingly good handle on breaking #news. Others decidedly don't.
#maduro #artificialintelligence #breakingnews
https://www.wired.com/story/us-invaded-venezuela-and-captured-nicolas-maduro-chatgpt-disagrees/
Show Original Post
hollywoodreporter (@hollywoodreporter@mastodon.social)
Beyond Cheating: AI at Hollywood’s Top Private Schools
#Lifestyle #LifestyleNews #ArtificalIntelligence #ChatGPT #GenerativeAI #Harvardwestlake #Privateschools
Show Original Post
BeginnersInAi (@BeginnersInAi@mastodon.social)
5️⃣ Here's the original story:
https://www.digitaltrends.com/computing/even-chatgpt-gets-anxiety-so-researchers-gave-it-a-dose-of-mindfulness-to-calm-down/
#AI #ChatGPT #ArtificialIntelligence #MachineLearning #Tech #AIEthics
Show Original Post
p (@p@pixelfed.social)
It worked, it did indeed make me feel better 😊 #ai #artificialintelligence #chatgpt #chatbot #linux #unix #nerd #geek #computer #tech
Show Original Post
rmblaber1956 (@rmblaber1956@mastodon.social)
https://www.theguardian.com/technology/2026/jan/02/google-ai-overviews-risk-harm-misleading-health-information. Again, why am I so not surprised? It isn't just #ChatGPT that's harmful to vulnerable people (especially young people) with #mentalhealth issues. Just how many of them have to die before action is taken?
Show Original Post
ThePhoenixVents (@ThePhoenixVents@todon.eu)
AI companies need to be held accountable for the lives they've ended and ruined.
#ChatGPT #AI
https://youtu.be/mlCgxxtct4Y
Show Original Post
habr (@habr@zhub.link)
Как защищаться от ИИ-психоза: гайд даже для тех кто думает что их это не коснется
Я провёл неглубокое исследование научных публикаций 2023-2025 годов о влиянии ИИ-чатботов на психическое здоровье. Спойлер: влияние есть, но доказательная база пока слабая. Зато практические рекомендации по защите -> вполне конкретные.
https://habr.com/ru/articles/982598/
#психоз #психология #ии #ииагенты #ии_чатбот #ииассистент #инкубаторы #chatgpt #gemini #claude
Show Original Post