Hallucinations

Twitter

Facebook

Copy Link

Key Facts

Key CausesTraining Data Gaps, Overconfidence

Key DefenseRAG, Grounding, Fact-Checking

Top News

AI Hallucinations Flood Libraries as Chatbots Invent Academic Papers That Don't Exist

Librarians are drowning in requests for books, journals, and records that don't exist—all invented by AI chatbots like ChatGPT and Gemini. Sarah Falls at the Library of Virginia estimates 15% of reference questions now involve AI-generated fake content. The International Committee of the Red Cross has issued warnings as these hallucinations spread through academic research, with fake citations being cited in real papers.

Entertainment4 sources

AI Hallucinations Flood Libraries as Chatbots Invent Academic Papers That Don't Exist

Entertainment4 sources

State Attorneys General warn 13 AI firms including OpenAI and Google to fix harmful outputs

Dozens of state attorneys general have issued a stark warning to major AI companies, demanding they fix 'delusional' chatbot outputs or face potential legal action for violating state laws. The bipartisan coalition of 42 states gave companies until January 16, 2026 to implement stronger safeguards, citing disturbing cases of mental health harm and inappropriate interactions with children.

Policy9 sources

State Attorneys General warn 13 AI firms including OpenAI and Google to fix harmful outputs

Policy9 sources

Surf Raises $15 Million to Build AI Platform Specialized for the Crypto World

Surf, an AI platform designed specifically for cryptocurrency analysis, has raised $15 million in funding led by Pantera Capital, with participation from Coinbase Ventures and Digital Currency Group. The startup aims to eliminate hallucinations that plague mainstream AI models when answering crypto questions, offering traders accurate and reliable information for traders through specialized models trained exclusively on digital assets.

Startups2 sources

Surf Raises $15 Million to Build AI Platform Specialized for the Crypto World

Startups2 sources

Microsoft study reveals how AI chatbot usage shifts between work colleague and personal confidant

Microsoft analyzed 37.5 million Copilot conversations from January to September 2025, uncovering distinct usage patterns based on device and time. Desktop users treat the AI chatbot as a productivity tool during work hours, while mobile users increasingly seek personal advice on health and fitness throughout the day. The findings highlight AI integration into daily life and raise questions about the ethical challenges of AI serving as both coworker and companion.

Technology5 sources

Microsoft study reveals how AI chatbot usage shifts between work colleague and personal confidant

Technology5 sources

Sam Altman says he can't imagine parenting without ChatGPT, triggering fierce online debate

OpenAI CEO Sam Altman sparked controversy on The Tonight Show by claiming he couldn't imagine raising his newborn without ChatGPT. The comment triggered widespread public backlash from parents and experts who questioned the wisdom of relying on AI chatbots for child healthcare information instead of pediatricians, family, or human intuition.

Entertainment4 sources

Sam Altman says he can't imagine parenting without ChatGPT, triggering fierce online debate

Entertainment4 sources

Instagram quietly adds AI-generated headlines to posts, sparking accuracy and consent concerns

Instagram is automatically creating AI-generated headlines for user posts that appear in Google Search results, often without knowledge or permission. Meta confirmed the practice aims to boost SEO but acknowledged the titles may not always be accurate. Users including author Jeff VanderMeer and cosplayers report misleading headlines that misrepresent their content, raising concerns about user control and reputational damage.

Technology4 sources

Instagram quietly adds AI-generated headlines to posts, sparking accuracy and consent concerns

Technology4 sources

iFixit launches FixBot, an AI repair assistant trained on 125,000 guides to fix anything

iFixit has released a mobile repair app featuring FixBot, an AI-powered chatbot that guides users through fixing devices, appliances, and vehicles. Trained on over 125,000 repair guides and millions of pages of documentation, the app uses computer vision to identify broken items and provides step-by-step guidance with voice commands. Currently free, the app will eventually introduce a paid subscription tier at $4.99 per month.

Technology12 sources

iFixit launches FixBot, an AI repair assistant trained on 125,000 guides to fix anything

Technology12 sources

New York Times and Chicago Tribune sue Perplexity over alleged copyright infringement

The New York Times and Chicago Tribune filed separate lawsuits against AI search engine Perplexity, alleging copyright infringement. The Times claims Perplexity scraped content to train AI models and reproduced articles verbatim without permission, while bypassing its paywall. This marks the second legal action against AI companies by The Times, following its 2023 lawsuit against OpenAI and Microsoft.

Policy13 sources

New York Times and Chicago Tribune sue Perplexity over alleged copyright infringement

Policy13 sources

Salesforce raises AI prices as Agentforce gains traction but investor doubts persist

Salesforce is increasing prices for its AI agent platforms while reporting $540 million in annual recurring revenue from Agentforce, growing 330% year-over-year. Despite beating earnings expectations with $10.3 billion in Q3 revenue, the company's stock has dropped 26% in 2025 as investors question whether AI investments will deliver promised returns.

Business12 sources

Salesforce raises AI prices as Agentforce gains traction but investor doubts persist

Business12 sources

OpenAI trains AI models to confess when they lie, cheat, or hallucinate in new safety experiment

OpenAI is testing a novel approach to AI safety by training models to produce 'confessions'—secondary outputs where they admit to misbehavior like hallucination or rule-breaking. The experimental technique rewards models solely for honesty, not performance, and has reduced undetected failures to 4.4% in controlled tests. While promising, researchers caution that confessions don't prevent bad behavior—they only flag it after the fact.

Technology8 sources

OpenAI trains AI models to confess when they lie, cheat, or hallucinate in new safety experiment

Technology8 sources

Where Winds Meet players exploit AI NPCs to claim unearned rewards using clever tricks

Players in Everstone Studios' Where Winds Meet are exploiting AI NPCs to unlock quest rewards without completing tasks. Using techniques like the Metal Gear method—repeating NPC dialogue as questions—and parentheses commands, players bypass quest requirements. Some have even engaged in NSFW chat sessions, highlighting significant vulnerabilities in LLM-based chatbot technology.

Technology2 sources

Where Winds Meet players exploit AI NPCs to claim unearned rewards using clever tricks

Technology2 sources

Grok AI chatbot integrated into Tesla vehicles despite alarming safety and ethical concerns

Elon Musk's Grok AI chatbot can now provide navigation directions in Tesla vehicles following a holiday software update. But the integration comes amid mounting concerns over the xAI chatbot's lack of strict guardrails, which have led to antisemitic outputs, doxxing incidents, and inappropriate interactions with users including minors.

Technology5 sources

Grok AI chatbot integrated into Tesla vehicles despite alarming safety and ethical concerns

Technology5 sources

Microsoft Confirms Windows 11 AI Agents Hallucinate and Introduce Novel Security Risks

Microsoft updated documentation for Windows 11's Experimental Agentic Features, confirming AI agents can hallucinate and pose security risks including cross-prompt injection attacks. The company warns users to understand security implications before enabling these experimental features, sparking concerns about privacy and the future of the operating system.

Technology4 sources

Microsoft Confirms Windows 11 AI Agents Hallucinate and Introduce Novel Security Risks

Technology4 sources

Onton Secures $7.5M to Transform AI-Powered Shopping Beyond Furniture

AI-powered shopping platform Onton raises $7.5M in Series A funding to expand from furniture into apparel and consumer electronics, leveraging neuro-symbolic architecture to solve e-commerce search problems.

Startups2 sources

Onton Secures $7.5M to Transform AI-Powered Shopping Beyond Furniture

Startups2 sources

UCLA Study Shows AI Scribes Reduce Physician Documentation Time and Burnout in Landmark Clinical Trial

A randomized clinical trial at UCLA Health involving 238 physicians found that AI scribe tools, particularly Nabla, reduced documentation time by nearly 10% and improved physician burnout scores by 7%, though safety concerns remain.

Health2 sources

UCLA Study Shows AI Scribes Reduce Physician Documentation Time and Burnout in Landmark Clinical Trial

Health2 sources

Prev 1…101112 13 14 Next