HorXi
horxi.bsky.social
HorXi
@horxi.bsky.social
Interested in Information Security, AI, Strategic Learning, Martial Arts.
Ein französischer Richter erlebt, was passieren kann, wenn die US-Regierung Tech-Firmen als Waffe einsetzt: ein Leben ohne Hotelreservierungen, Amazon-Shopping, Paypal und Kreditkarten.

www.golem.de/news/interna...
Golem.de: IT-News für Profis
www.golem.de
November 28, 2025 at 8:01 PM
Forscher entdecken massive Sicherheitslücke, bei der einfache Reime die Schutzmechanismen von Google und OpenAI komplett aushebeln.

www.all-ai.de/news/topbeit...
Goethe schlägt GPT-5: Warum Gedichte jede KI knacken
Forscher entdecken massive Sicherheitslücke, bei der einfache Reime die Schutzmechanismen von Google und OpenAI komplett aushebeln.
www.all-ai.de
November 28, 2025 at 7:35 PM
Data regulators in Thailand said they are blocking the Sam Altman-founded company Tools for Humanity from collecting citizens’ iris scans in exchange for cryptocurrency payments.

therecord.media/thailand-wor...
Thailand bans World iris scans, orders company to delete data
Data regulators in Thailand said they are blocking the Sam Altman-founded company Tools for Humanity from collecting iris scans in exchange for cryptocurrency payments.
therecord.media
November 28, 2025 at 6:15 AM
An emergency warning service used by municipalities across the U.S. was taken down in recent weeks by hackers who also stole information on the system's users.

therecord.media/emergency-wa...
Municipal emergency warning service offline after hackers steal user data
An emergency warning service used by municipalities across the U.S. was taken down in recent weeks by hackers who also stole information on the system's users.
therecord.media
November 28, 2025 at 6:14 AM
Researchers affiliated with Carnegie Mellon University, Instituto Superior Técnico/INESC-ID, and AI security platform Hydrox AI describe their approach in a preprint paper titled "RECAP: Reproducing Copyrighted Data from LLMs Training with an Agentic Pipeline."

www.theregister.com/2025/11/21/r...
Researchers build a better AI model memory probe
: RECAP agent overcomes model alignment efforts to hide memorized proprietary content
www.theregister.com
November 28, 2025 at 6:12 AM
Computer scientists based in South Korea have devised what they describe as an "AI Kill Switch" to prevent AI agents from carrying out malicious data scraping.

www.theregister.com/2025/11/21/b...
Boffins build 'AI Kill Switch' to thwart unwanted agents
: AutoGuard uses injection text for good
www.theregister.com
November 28, 2025 at 6:09 AM
At New Zealand's Kawaiicon cybersecurity convention, organizers hacked together a way for attendees to track CO2 levels throughout the venue—even before they arrived.

www.wired.com/story/this-h...
This Hacker Conference Installed a Literal Antivirus Monitoring System
At New Zealand's Kawaiicon cybersecurity convention, organizers hacked together a way for attendees to track CO2 levels throughout the venue—even before they arrived.
www.wired.com
November 28, 2025 at 6:08 AM
The deepfake fraud wave has arrived, and its most dangerous hour might surprise you…

cybernews.com/cybercrime/a...
AI-powered fraud is exploding: it peaks in a very specific two-hour window
The deepfake fraud wave has arrived, and its most dangerous hour might surprise you.
cybernews.com
November 27, 2025 at 5:50 AM
Despite growing evidence that AI is displacing workers, new research reveals a striking disconnect: while people acknowledge automation will transform the job market, they remain stubbornly confident it won't affect them personally.

cybernews.com/ai-news/are-...
Are people underestimating the chances that AI will take their jobs?
Despite growing evidence that AI is displacing workers, new research reveals a striking disconnect.
cybernews.com
November 27, 2025 at 5:48 AM
The UK Ministry of Defence has reportedly placed warnings in electric cars from China, banning staff and troops from discussing sensitive information amid spying fears.

cybernews.com/news/uk-chin...
UK MoD warns staff not to discuss secrets in cars from China
The UK Ministry of Defence has warned its staff and troops against discussing sensitive information in electric vehicles manufactured in China.
cybernews.com
November 27, 2025 at 5:46 AM
Businesses will soon commercialize quantum computing, but nation states might weaponize the technology even sooner, Palo Alto Networks CEO Nikesh Arora has suggested.

cybernews.com/security/pal...
Palo Alto boss says nation states will soon weaponize quantum computing
Anticipating a rush of quantum-related cyber dangers, Arora expects his cybersecurity firm to use the opportunity, offer a bunch of quantum-safe products, and cash in.
cybernews.com
November 27, 2025 at 5:45 AM
Endpoint security is often overlooked, costing companies millions in revenue when recovering from a cyberattack. By creating an endpoint environment that can self-heal, enterprises can maintain operations even when primary systems are compromised.

cybernews.com/security/why...
Exclusive interview with IGEL's Jason Mafera. Why endpoint security is still the weakest link in cyber defense
Endpoint security is often overlooked, costing companies millions in revenue when recovering from a cyberattack.
cybernews.com
November 27, 2025 at 5:43 AM
Information proliferation in the internet age has government auditors reiterating that loose tweets can sink fleets, and they're concerned that the Defense Department isn't doing enough to stop sensitive info from getting out there.

www.theregister.com/2025/11/17/p...
Pentagon and soldiers let too many secrets slip on socials
: Ready, aim, mire
www.theregister.com
November 26, 2025 at 6:10 AM
IoT devices can be compromised, thanks to gaps in cloud management interfaces for firewalls and routers, even if they're protected by security software or not online.

www.darkreading.com/cybersecurit...
IoT Devices Open to Silent Takeover Via Cloud Firewalls
IoT devices can potentially be compromised in droves, thanks to gaps in cloud management interfaces for firewalls and routers, even if they're not online.
www.darkreading.com
November 26, 2025 at 6:08 AM
Two of the most populous states have moved to legislate controls on workplace surveillance amid a boom in work from home and artificial intelligence, sparking concerns from employers of overreach.

www.afr.com/work-and-car...
States to tighten rules on watching workers in age of WFH and AI
The Allan government has backed greater protections for workers against spying bosses as technology and work from home test the boundaries of employee privacy.
www.afr.com
November 26, 2025 at 6:06 AM
Two British teenagers charged with Computer Misuse Act offenses over a cyberattack on Transport for London (TfL) last year pleaded not guilty during a court appearance on Friday.

therecord.media/transport-fo...
Two suspected Scattered Spider hackers plead not guilty over Transport for London cyberattack
Two U.K. teenagers pleaded not guilty to hacking the Transport for London agency in 2024 — an attack attributed to the Scattered Spider cybercrime group.
therecord.media
November 25, 2025 at 6:02 AM
Police departments across the country searched records from a national network of automated license plate reader cameras hundreds of times over the last year to track protest activity, according to new research from the Electronic Frontier Foundation (EFF).

therecord.media/flock-safety...
Flock Safety cameras used to monitor protesters, rights group finds
Police departments across the country searched records from Flock Safety's network of automated license plate reader cameras hundreds of times to track protest activity, researchers found.
therecord.media
November 25, 2025 at 6:01 AM
The China-linked hacking group known as APT31 infiltrated Russia’s technology sector for years and quietly exfiltrated data from companies involved in government contracting and systems integration, according to a new report.

therecord.media/russia-repor...
China’s APT31 linked to hacks on Russian tech firms
Moscow-based Positive Technologies says a China-linked group tracked as APT31 appears to be responsible for breaches of entities in Russia's tech sector.
therecord.media
November 25, 2025 at 5:59 AM
Security researchers have uncovered a new Android banking trojan capable of intercepting messages from apps including WhatsApp, Telegram and Signal after they have been decrypted.

therecord.media/new-android-...
New Android malware can capture private messages, researchers warn
The malware can monitor everything displayed on a phone in real time — including contacts, full message threads and the content of encrypted chats — by accessing data after it has been decrypted by le...
therecord.media
November 24, 2025 at 6:07 AM
Cloud giant Salesforce warned customers of a potential data breach on Wednesday evening after discovering “unusual activity” related to a third-party application called Gainsight.

therecord.media/salesforce-c...
Salesforce cuts off access to third-party app after discovering ‘unusual activity’
Salesforce posted a message on its website saying an investigation revealed that the activity “may have enabled unauthorized access to certain customers’ Salesforce data through the app’s connection.”
therecord.media
November 24, 2025 at 6:05 AM
A teddy bear telling a child about sexual kinks or where to find knives is no longer science fiction but the sad reality of toys that integrate chatbots, a report has found.

cybernews.com/ai-news/ai-t...
AI toys talk about sex with children, give advice on finding knives
A new report sheds light on the dangers of AI-powered toys – some have been found to engage in adult conversations and instruct children on where to find sharp objects.
cybernews.com
November 24, 2025 at 6:03 AM
A trio of autumn outages in a four-week period highlights how configuration and metadata errors in the cloud are becoming “the new power cuts.”

cybernews.com/cloud/cloudf...
Cloudflare, Azure, AWS: the striking pattern behind the outage cluster shaking the internet
A trio of autumn outages in a four-week period highlights how configuration and metadata errors in the cloud are becoming “the new power cuts.”
cybernews.com
November 24, 2025 at 6:02 AM
As government regulation lags, some insurance companies see a business case for pushing AI companies to minimize risk and adopt stronger guardrails.

www.nbcnews.com/tech/tech-ne...
Insurance companies are trying to avoid big payouts by making AI safer
As government regulation lags, some insurance companies see a business case for pushing AI companies to minimize risk and adopt stronger guardrails.
www.nbcnews.com
November 23, 2025 at 11:49 AM
Microsoft on Monday disclosed that it automatically detected and neutralized a distributed denial-of-service (DDoS) attack targeting a single endpoint in Australia that measured 15.72 terabits per second (Tbps) and nearly 3.64 billion packets per second (pps).

thehackernews.com/2025/11/micr...
Microsoft Mitigates Record 15.72 Tbps DDoS Attack Driven by AISURU Botnet
Microsoft reports a record 5.72 Tbps AISURU DDoS attack as related IoT botnets continue evolving.
thehackernews.com
November 23, 2025 at 11:47 AM
EU plans to simplify AI and privacy laws
Proposals still face debate, votes in region
Critics argue changes favour Big Tech, Trump
Biggest digital rights rollback in EU history, activists say

www.reuters.com/sustainabili...
www.reuters.com
November 23, 2025 at 11:46 AM