Skip to main content

Posts

Showing posts from 2024

Hacker plants false memories in ChatGPT to steal user data in perpetuity

Emails, documents, and other untrusted content can plant malicious memories. https://arstechnica.com/?p=2052198&utm_source=dlvr.it&utm_medium=blogger

Terminator’s Cameron joins AI company behind controversial image generator

Famed sci-fi director joins board of embattled Stability AI, creator of Stable Diffusion. https://arstechnica.com/?p=2052114&utm_source=dlvr.it&utm_medium=blogger

Broadcom responds to AT&T’s VMware support lawsuit: AT&T has “other options”

Broadcom defends against renewal, citing "End of Availability" provision. https://arstechnica.com/?p=2052026&utm_source=dlvr.it&utm_medium=blogger

OpenAI CEO: We may have AI superintelligence in “a few thousand days”

Altman says "deep learning worked" and will lead to "massive prosperity." https://arstechnica.com/?p=2051754&utm_source=dlvr.it&utm_medium=blogger

11 million devices infected with botnet malware hosted in Google Play

Necro infiltrated Google Play in 2019. It recently returned. https://arstechnica.com/?p=2051854&utm_source=dlvr.it&utm_medium=blogger

When you call a restaurant, you might be chatting with an AI host

Voice chatbots are increasingly picking up the phone for restaurants. https://arstechnica.com/?p=2051613&utm_source=dlvr.it&utm_medium=blogger

Secret calculator hack brings ChatGPT to the TI-84, enabling easy cheating

Tiny device installed inside TI-84 enables Wi-Fi Internet, access to AI chatbot. https://arstechnica.com/?p=2051342&utm_source=dlvr.it&utm_medium=blogger

Google calls for halting use of WHOIS for TLS domain verifications

WHOIS data is unreliable. So why is it used in TLS certificate applications? https://arstechnica.com/?p=2051511&utm_source=dlvr.it&utm_medium=blogger

NC governor candidate cries AI fabrication as defense for racist porn forum posts

Mark Robinson's claims of AI-generated slander show that the "deep doubt era" is upon us. https://arstechnica.com/?p=2051248&utm_source=dlvr.it&utm_medium=blogger

Microsoft releases a new Windows app called Windows App for running Windows apps

Windows App replaces Microsoft Remote Desktop on macOS, iOS, and Android. https://arstechnica.com/?p=2051069&utm_source=dlvr.it&utm_medium=blogger

Ever wonder how crooks get the credentials to unlock stolen phones?

iServer provided a simple service for phishing credentials to unlock phones. https://arstechnica.com/?p=2051165&utm_source=dlvr.it&utm_medium=blogger

Real-time Linux is officially part of the kernel after decades of debate

Now you can run your space laser or audio production without specialty patches. https://arstechnica.com/?p=2050956&utm_source=dlvr.it&utm_medium=blogger

$100 billion AI infrastructure fund launched by Microsoft, BlackRock, UAE firm

Investment "goes beyond what any single company or government can finance." https://arstechnica.com/?p=2050936&utm_source=dlvr.it&utm_medium=blogger

“Dead Internet theory” comes to life with new AI-powered social media app

SocialAI takes the social media "filter bubble" to an extreme with 100% fake interactions. https://arstechnica.com/?p=2050587&utm_source=dlvr.it&utm_medium=blogger

Landmark AI deal sees Hollywood giant Lionsgate provide library for AI training

Runway deal will create a Lionsgate AI video generator used to augment filmmaking. https://arstechnica.com/?p=2050685&utm_source=dlvr.it&utm_medium=blogger

Massive China-state IoT botnet went undetected for four years—until now

75% of infected devices were located in homes and offices in North America and Europe. https://arstechnica.com/?p=2050629&utm_source=dlvr.it&utm_medium=blogger

Due to AI fakes, the “deep doubt” era is here

As AI deepfakes sow doubt in legitimate media, anyone can claim something didn't happen. https://arstechnica.com/?p=2042584&utm_source=dlvr.it&utm_medium=blogger

Google seeks authenticity in the age of AI with new content labeling system

C2PA system aims to give context to search results, but trust problems run deeper than AI tech. https://arstechnica.com/?p=2050363&utm_source=dlvr.it&utm_medium=blogger

Ban warnings fly as users dare to probe the “thoughts” of OpenAI’s latest model

OpenAI does not want anyone to know what o1 is “thinking" under the hood. https://arstechnica.com/?p=2049959&utm_source=dlvr.it&utm_medium=blogger

Secure Boot-neutering PKfail debacle is more prevalent than anyone knew

Keys were marked "DO NOT TRUST." More devices than previously known used them anyway. https://arstechnica.com/?p=2050182&utm_source=dlvr.it&utm_medium=blogger

Omnipresent AI cameras will ensure good behavior, says Larry Ellison

"We’re going to have supervision," says billionaire Oracle co-founder Ellison. https://arstechnica.com/?p=2049906&utm_source=dlvr.it&utm_medium=blogger

1.3 million Android-based TV boxes backdoored; researchers still don’t know how

Infection corrals devices running AOSP-based firmware into a botnet. https://arstechnica.com/?p=2049773&utm_source=dlvr.it&utm_medium=blogger

Google rolls out voice-powered AI chat to the Android masses

Gemini Live allows back-and-forth conversation, now free to all Android users. https://arstechnica.com/?p=2049707&utm_source=dlvr.it&utm_medium=blogger

Free Starlink Internet is coming to all of United’s airplanes

The upgrade starts in 2025 but with more than 1,000 planes, will take several years. https://arstechnica.com/?p=2049628&utm_source=dlvr.it&utm_medium=blogger

OpenAI’s new “reasoning” AI models are here: o1-preview and o1-mini

New o1 language model can solve complex tasks, count Rs in "strawberry." https://arstechnica.com/?p=2049445&utm_source=dlvr.it&utm_medium=blogger

Music industry’s 1990s hard drives, like all HDDs, are dying

The music industry traded tape for hard drives and got a hard-earned lesson. https://arstechnica.com/?p=2049391&utm_source=dlvr.it&utm_medium=blogger

My dead father is “writing” me notes again

A recent AI discovery resurrected my late father's handwriting—and I want anyone to use it. https://arstechnica.com/?p=2045545&utm_source=dlvr.it&utm_medium=blogger

As quantum computing threats loom, Microsoft updates its core crypto library

Two algorithms added so far, two more planned in the coming months. https://arstechnica.com/?p=2049244&utm_source=dlvr.it&utm_medium=blogger

Taylor Swift cites AI deepfakes in endorsement for Kamala Harris

Taylor Swift on AI: "The simplest way to combat misinformation is with the truth." https://arstechnica.com/?p=2049045&utm_source=dlvr.it&utm_medium=blogger

Rogue WHOIS server gives researcher superpowers no one should ever have

.mobi top-level-domain managers changed the location of its WHOIS server. No one got the memo. https://arstechnica.com/?p=2048683&utm_source=dlvr.it&utm_medium=blogger

Roblox announces AI tool for generating 3D game worlds from text

New AI feature aims to streamline game creation on popular online platform. https://arstechnica.com/?p=2048533&utm_source=dlvr.it&utm_medium=blogger

Found: 280 Android apps that use OCR to steal cryptocurrency credentials

Optical Character Recognition converts passwords shown in images to machine-readable text. https://arstechnica.com/?p=2047970&utm_source=dlvr.it&utm_medium=blogger

Nvidia’s AI chips are cheaper to rent in China than US

Supply of processors helps Chinese startups advance AI technology despite US restrictions. https://arstechnica.com/?p=2047945&utm_source=dlvr.it&utm_medium=blogger

WaveCore runs right through a concrete wall with gigabit-speed network signal

Core drilling is tricky. Getting a 6 GHz signal through concrete is now easier. https://arstechnica.com/?p=2047825&utm_source=dlvr.it&utm_medium=blogger

AT&T sues Broadcom for refusing to renew perpetual license support

Ars cited in lawsuit AT&T recently filed against Broadcom. https://arstechnica.com/?p=2047618&utm_source=dlvr.it&utm_medium=blogger

FBI busts musician’s elaborate AI-powered $10M streaming-royalty heist

Feds say it's the first US criminal case involving artificially inflated music streaming. https://arstechnica.com/?p=2047451&utm_source=dlvr.it&utm_medium=blogger

Generative AI backlash hits annual writing event, prompting resignations

NaNoWriMo refuses to condemn AI as accessibility tool, faces criticism from writers. https://arstechnica.com/?p=2047101&utm_source=dlvr.it&utm_medium=blogger

Zyxel warns of vulnerabilities in a wide range of its products

Most serious vulnerabilities carry severity ratings of 9.8 and 8.1 out of a possible 10. https://arstechnica.com/?p=2047312&utm_source=dlvr.it&utm_medium=blogger

Sutskever strikes AI gold with billion-dollar backing for superintelligent AI

Top venture firms back SSI to develop "safe" AI with teams in Palo Alto and Tel Aviv. https://arstechnica.com/?p=2047175&utm_source=dlvr.it&utm_medium=blogger

YubiKeys are vulnerable to cloning attacks thanks to newly discovered side channel

Sophisticated attack breaks security assurances of the most popular FIDO key. https://arstechnica.com/?p=2046777&utm_source=dlvr.it&utm_medium=blogger

City of Columbus sues man after he discloses severity of ransomware attack

Mayor said data was unusable to criminals; researcher proved otherwise. https://arstechnica.com/?p=2046614&utm_source=dlvr.it&utm_medium=blogger

ChatGPT hits 200 million active weekly users, but how many will admit using it?

Despite corporate prohibitions on AI use, people flock to the chatbot in record numbers. https://arstechnica.com/?p=2046413&utm_source=dlvr.it&utm_medium=blogger

Commercial spyware vendor exploits used by Kremlin-backed hackers, Google says

Findings undercut pledges of NSO Group and Intgellexa their wares won't be abused. https://arstechnica.com/?p=2046289&utm_source=dlvr.it&utm_medium=blogger

Unpatchable 0-day in surveillance cam is being exploited to install Mirai

Vulnerability is easy to exploit and allows attackers to remotely execute commands. https://arstechnica.com/?p=2046043&utm_source=dlvr.it&utm_medium=blogger

New AI model can hallucinate a game of 1993’s Doom in real time

Dobos: "Why write rules for software by hand when AI can just think every pixel for you?" https://arstechnica.com/?p=2045767&utm_source=dlvr.it&utm_medium=blogger

“Exploitative” IT firm has been delaying 2,000 recruits’ onboarding for years

India's Infosys recruits reportedly subjected to repeated, unpaid "pre-training." https://arstechnica.com/?p=2045697&utm_source=dlvr.it&utm_medium=blogger

Debate over “open source AI” term brings new push to formalize definition

Restrictive AI model licenses claimed as "open source" spark for clear standard. https://arstechnica.com/?p=2045225&utm_source=dlvr.it&utm_medium=blogger

Hackers infect ISPs with malware that steals customers’ credentials

Zero-day that was exploited since June to infect ISPs finally gets fixed. https://arstechnica.com/?p=2045401&utm_source=dlvr.it&utm_medium=blogger

Hobbyists discover how to insert custom fonts into AI-generated images

Like adding custom art styles or characters, in-world typefaces come to Flux. https://arstechnica.com/?p=2045327&utm_source=dlvr.it&utm_medium=blogger

Android malware steals payment card data using previously unseen technique

Attacker then emulates the card and makes withdrawals or payments from victim's account. https://arstechnica.com/?p=2045086&utm_source=dlvr.it&utm_medium=blogger