All Categories
AI Bots in Video Calls Now Criticize How Employees Speak
Companies are increasingly using AI bots in video meetings to transcribe, summarize, and even critique participants' speaking style. While some find the feedback helpful, others describe the experience as eerie or dystopian. The trend raises questions about workplace surveillance and the role of AI in communication.
Search Giant Keeps Suicide-Encouraging Forum in Results
A New York Times investigation reveals a website where users encourage suicide, linked to at least 45 deaths. Google, citing open access to information, has not removed the site from its search results, prompting debate over tech companies' responsibility in moderating harmful content.
AI Safety Gaps Could Turn Catastrophic Within a Decade, Ex-Google Chief Says
Eric Schmidt, former Google CEO and current chair of the National Security Commission on Artificial Intelligence, told an Axios summit that current AI safeguards are insufficient and that the technology could become dangerous within five to ten years. He proposed an IPCC-style body to guide policymakers. Meanwhile, Meta's Yann LeCun disputes the urgency, arguing that AI is far from human-level threat.
Methane Turbines Roar in Mississippi as AI Data Center Sparks Local Outcry
Residents of Southaven, Mississippi, are facing relentless noise and air pollution from a $20 billion xAI data center powered by 27 methane turbines. Locals report health issues and disrupted lives, while the company seeks permits for 41 permanent turbines. Officials acknowledge concerns but face skepticism over mitigation efforts.
AI-Generated Fake News Finds a Home on Reddit
A new Reddit community, newsbyAI, showcases fabricated news articles written by AI systems like Grover and GPT-2. The posts range from celebrity gossip to corporate scandals, highlighting the potential for AI to spread convincing misinformation online.
Data-Driven Algorithm Aims to Boost Refugee Employment Through Smarter Placement
Researchers from Stanford University and ETH Zurich have developed an algorithm that predicts optimal refugee placement locations to improve employment outcomes. Testing on historical data from the U.S. and Switzerland suggests significant employment gains could be achieved without additional government costs. The team emphasizes human oversight remains central to the process.
AI Stock Frenzy Faces Investor Doubts as Funds Trim Tech Bets
As concerns grow over frothy AI valuations, some major investment funds are reducing exposure to big tech, while others see continued gains. The debate highlights uncertainty over when, or if, a market correction may occur.
Box Elder County's 40,000-Acre Data Center Plan Draws Fire
A proposed 40,000-acre data center in Box Elder County, Utah, backed by TV personality Kevin O'Leary, has ignited fierce local opposition. Residents cite environmental concerns for the shrinking Great Salt Lake, rising utility costs, and a lack of transparency in the approval process, prompting threats of legal action.
OpenAI Warns of Fraudulent Equity Schemes Targeting AI Investors
OpenAI has issued a public warning about unauthorized offers to buy its equity, cautioning investors against fraudulent schemes that could result in financial loss and legal liability. The company emphasizes that any transfer without its written consent is void and may violate securities laws.
OpenAI CEO and AI Critic Clash Over Hype and Reality
A public spat between OpenAI CEO Sam Altman and cognitive scientist Gary Marcus highlights a growing rift over AI's promises versus its actual capabilities. Marcus accused Altman of overhyping AI, while Altman defended his company's achievements, underscoring a broader industry tension.
Bots Outperform People in CAPTCHA Challenges, Study Finds
A new study from UC Irvine reveals that AI-powered bots solve CAPTCHA puzzles with 99.8% accuracy, far surpassing humans who range from 50-84%. The research, involving 1,400 participants and 14,000 tests, questions the effectiveness of these widely used security measures.
New AI Model Decodes Keystrokes from Zoom Audio with High Accuracy
A UK research team has developed an AI model that can identify what keys are pressed on a laptop keyboard with up to 95% accuracy using only audio from a nearby microphone or Zoom call. The study, not yet peer-reviewed, highlights growing cybersecurity risks from ubiquitous microphones and deep learning. Researchers suggest using complex passwords and two-factor authentication to mitigate potential threats.
CIA Builds AI Chatbots to Simulate Foreign Leaders' Reactions
The CIA has developed a platform using large language models to let analysts interact with simulated versions of foreign leaders, aiming to predict their decisions. The initiative, led by the agency's first CTO, seeks to modernize intelligence work and streamline private-sector partnerships, though details remain classified.
Purdue Study: ChatGPT Gives Wrong Coding Answers More Than Half the Time
A Purdue University study presented at a human-computer interaction conference reveals that 52% of ChatGPT's answers to programming questions contain misinformation, and 77% are more verbose than human responses. The research, based on 517 Stack Overflow questions, highlights that programmers often prefer ChatGPT's polite tone, which can mask errors.
DeepSeek's Next AI Model Poised to Test US Tech Market Again
DeepSeek is set to release its V4 AI model, a move that could rattle US tech markets already strained by massive AI investments. The launch follows the V3 debut that caused a 3% Nasdaq drop and a $600 billion Nvidia loss. Investors and tech giants are bracing for potential market volatility.
Google’s Bard Chatbot Backs DOJ in Antitrust Case, Calls for Breakup
Days after its public release, Google's AI chatbot Bard sided with the Justice Department in an antitrust exchange, arguing the company holds a monopoly in digital ads. The bot's response, shared by a security researcher, highlights ongoing federal lawsuits against Google. This episode underscores the tension between AI tools and their corporate creators.
Bezos and Musk Beam Into Star Trek Pilot via Deepfake
A new deepfake video places Amazon's Jeff Bezos and SpaceX's Elon Musk into the original 1965 Star Trek pilot, with Bezos appearing as a Talosian alien. The clip highlights both the creative potential and the serious ethical concerns surrounding AI-generated likeness manipulation, including nonconsensual pornography and disinformation risks.
Writer Stands by AI Assistance After Book's False Quotes Exposed
Steven Rosenbaum, author of a book on AI and truth, faces backlash after The New York Times found fabricated quotes in his work, which he attributes to AI hallucinations. Despite the scandal, Rosenbaum defends his use of AI tools, calling them a 'delightful writing companion,' and says he won't return to traditional writing methods.
Lyft Rolls Out Autonomous Rides in Bay Area via Drive.ai Partnership
Lyft has begun offering self-driving car rides to select passengers in the San Francisco Bay Area, using technology from startup Drive.ai. The move aligns with Lyft's broader vision of a future dominated by autonomous fleets and a decline in personal car ownership, as outlined by co-founder John Zimmer.
Music Labels and Google Explore Licensing AI-Generated Vocals
Universal Music Group, after previously demanding removal of AI-cloned songs, is reportedly negotiating with Google to license artists' voices and melodies for fan-generated tracks. Warner Music is also involved, with executives suggesting such tools could offer new creative avenues while artists may opt out. The talks signal a potential shift toward monetizing AI deepfakes in the music industry.