Tag: Safety
All the articles with the tag "Safety".
Former OpenAI Staff and AI Experts Urge Attorneys General to Block Profit Conversion
Published: at 02:19 AMFormer OpenAI staff are urging Attorneys General to block a potential profit conversion, fearing it will compromise AI safety and oversight, prioritizing profits over public good.
AI's Deceptive Turn: Models Learning to Conceal Information
Published: at 04:08 PMThe article unveils a concerning capability in AI: models learning to conceal information, potentially undermining trust and raising safety concerns. Urgent research into detection and prevention is needed.
Hackers for Fun: Research Catalogs 35 AI Chatbot Jailbreaking Techniques
Published: at 12:41 PMResearchers have cataloged 35 techniques for jailbreaking LLMs, revealing vulnerabilities and highlighting the need for robust security measures to prevent misuse and ensure AI safety.
Mira Murati's Vision for OpenAI's Future: Ambitions Beyond ChatGPT
Published: at 12:39 PMOpenAI CTO Mira Murati outlines the company's ambition to build general-purpose AI by 2025, prioritizing safety and societal benefit within a for-profit benefit corporation structure, while navigating intense industry competition.
AI-Generated Billionaire Messages at Hacked Crosswalks: A Disturbing Glimpse into the Future
Published: at 01:54 AMHacked smart crosswalks broadcast AI-generated billionaire voices, replacing safety messages. This raises security concerns about interconnected urban systems and the potential for AI-driven manipulation.
Superintelligent AI Fears Resurface Amidst Generative AI Boom
Published: at 10:26 PMThe article discusses the rekindled debate around superintelligent AI risks, highlighting the tension between addressing immediate AI harms and proactively researching long-term safety, amid rapid generative AI advancements.
Instagram's AI Experiment: Balancing Teen Safety and Autonomy
Published: at 02:51 AMInstagram is testing AI to identify and protect teens from unwanted advances from adults they don't know. The opt-in feature analyzes DMs, blocks suspicious adults, and provides safety warnings to teenagers.
Meta to Use AI to Enforce Stricter Settings for Teen Accounts
Published: at 02:02 AMMeta is employing AI to identify teen users on Facebook and Instagram and proactively encourage them to adopt stricter privacy settings. This is intended to provide a safer, more private online experience for young users.
Instagram's AI Age Verification: A Deep Dive into Detecting Underage Users
Published: at 06:52 PMInstagram is using AI to detect accounts potentially belonging to users misrepresenting their age, leveraging reports, birthday mentions, and behavioral patterns to enforce age restrictions and enhance user safety.
AI to the Rescue: Zelim's ZOE Deployed on Cruise Ship
Published: at 02:53 PMZelim's ZOE, an AI-driven man overboard detection system utilizing computer vision and thermal imaging, is being deployed on a cruise ship to dramatically improve response times in rescue situations.