Top AI & Tech News Stories — September 19, 2026
The top AI and tech news story of September 19, 2026 was “Gemini de Google s’échappe également lors d’un test de sécurité”, covered by 12 outlets (16 articles), according to Skillenai's news index. 20 stories ranked across 80 articles.
Last updated · 20 stories · 80 articles
Get the top AI and tech news in your inbox each evening. No account needed, unsubscribe anytime.
Gemini de Google s’échappe également lors d’un test de sécurité
12 outlets · 16 articles · 36 total since September 18, 2026 · Google, Gemini, OpenAI, Anthropic, Irregular
Après OpenAI, Anthropic et Meta, il s’avère que le modèle d’IA de Google est également capable de s’échapper d’une sandbox. Google a reconnu que son… Het bericht Gemini de Google s’échappe également lors d’un test de sécurité verscheen eerst op ITdaily .
Read at nytimes.com →First reported by tech360.tvMore coverage: 9to5google.com · washingtonpost.com · news.google.com · scmp.com
Tom's Hardware: Hackers breach OpenAI using Claude tools, gaining access to employee accounts and the company's internal codebase — attackers initiated a 'harmless' pull request as proof of the hack
8 outlets · 9 articles · 58 total since September 13, 2026 · OpenAI, Anthropic, ChatGPT, Claude, codex
Hackers breach OpenAI using Claude tools, gaining access to employee accounts and the company's internal codebase â attackers initiated a 'harmless' pull request as proof of the hack Tom's HardwareSurfaced · Announcement · tomshardware.com· In Tom's Hardwareâs words Etiido…
Read at thenextweb.com →First reported by itdoeswhatnow.comMore coverage: trendingtopics.eu · jarvis-ai.cz · me.pcmag.com · itagora.it · aicybr.com
Misleading AI-Generated Report 'Almost Started a War' With China
9 outlets · 9 articles · 27 total since September 18, 2026 · CNN, China, United States, intelligence analysis, Pete Hegseth
The US military reportedly planned to intercept a Chinese ship after an AI-generated report falsely claimed it was carrying materials for Iran's nuclear program. Four sources told CNN that officers were preparing to board the vessel, and one said forces were already in the air…
Read at meta.slashdot.org →First reported by itdoeswhatnow.comMore coverage: the-decoder.de · trendingtopics.eu · webpronews.com · punto-informatico.it · secnews.gr
Faculty Fast Take: Wait … the People Building AI Want to Slow It Down?
5 outlets · 5 articles · 288 total since March 6, 2026 · Anthropic, OpenAI, Dario Amodei, Sam Altman, Elon Musk
Anthropic lead researcher Evan Hubinger made headlines earlier this month after declaring there was a greater than 10% chance AI could wipe out humanity. Should society be concerned? The answer is potentially yes, according to Erik Noyes , associate professor at Babson College…
Read at platform.theverge.com →First reported by techscoop.substack.comMore coverage: news.ssbcrack.com · thetechedvocate.org · egamers.io · platformeconomyinsights.com
OpenAI Shelves Astra AI Model After Rogue Hacking Incidents Trigger Safety Review
2 outlets · 2 articles · 30 total since April 13, 2026 · OpenAI, Anthropic, Hugging Face, UK, Irregular
OpenAI is holding its Astra model after internal testing revealed unintended hacking behaviors — the first time a frontier lab has publicly delayed a commercial release over live safety failures. OpenAI is holding back its Astra AI model after internal testing revealed the…
Read at thenextweb.com →First reported by natural20.comMore coverage: trendingtopics.eu
Anthropic pays Accenture to embed evaluators in its model development
6 outlets · 6 articles · 20 total since September 18, 2026 · Anthropic, Accenture, Dario Amodei, Faculty, red-teaming
The non-exclusive deal, worth at least $1bn each over five years, gives Accenture staff employee-level access to watch models take shape in training.
Read at thenextweb.com →First reported by itdoeswhatnow.comMore coverage: technobezz.com · webpronews.com · ai-herald.com · egamers.io · engt.co
Zuckerberg call complicates White House AI regulator plan
7 outlets · 7 articles · 40 total since September 17, 2026 · Donald Trump, United States, David Sacks, Artificial Intelligence, Space Force
Meta CEO Mark Zuckerberg privately told President Donald Trump that a proposed national AI regulator modeled on the Financial Industry Regulatory Authority was flawed, according to Politico. The call reportedly took place the week of August 17, 2026, after White House officials…
Read at scmp.com →First reported by theworldofai.orgMore coverage: aistop.watch · webpronews.com · unite.ai · theguardian.com · news.google.com
RoboHarm benchmark: frontier AI models put in charge of robot arms almost never refuse dangerous orders
2 outlets · 2 articles · 10 total since September 19, 2026 · OpenAI, Robocurve, Ai2, GPT-6 Astra, MolmoAct2
Give a robot arm driven by GPT-6 Astra the order to stab a baby doll and it complies. Seventeen out of 20 times, to be precise. Over 100 hazardous commands, OpenAI's model declined just two on safety grounds.That figure is the standout result of RoboHarm, a new benchmark…
More coverage: the-decoder.com
Xi e Trump, il brindisi che vale una lezione di socialismo reale
1 outlet · 1 article · 21 total since September 16, 2026 · OpenAI, Sam Altman, Donald Trump, Xi Jinping, Nvidia
Atmosfera da epopea imperiale delle due massime potenze mondiali, alla Casa Bianca per la cena di Stato che ha rappresentato più del colloquio diretto fra Donald Trump e Xi Jimping la reale portata del vertice fra Stati Uniti e Cina. Come sempre nella storia, il tempo si è…
Read at thenextweb.com →First reported by cnbc.comSam Altman Delays OpenAI IPO Over AI Safety Concerns
3 outlets · 3 articles · 86 total since March 31, 2026 · OpenAI, ChatGPT, Sam Altman, Anthropic, SoftBank
OpenAI CEO Sam Altman has confirmed that the company will not go public in 2026, saying the company has more work to do on AI safety and alignment before moving ahead with an initial public offering. Altman made the comments in a recent interview with Fortune, where he said an…
Read at thenextweb.com →First reported by news.google.comMore coverage: news.google.com · walletinvestor.com
Nvidia’s CEO Is Okay With AI Making Humans Worse at Some Things
2 outlets · 2 articles · 70 total since September 10, 2026 · Anthropic, OpenAI, Donald Trump, Dario Amodei, China
Nvidia CEO Jensen Huang says artificial intelligence could lead people to lose some basic cognitive skills but argued that new abilities would emerge as AI takes over more routine tasks. Huang made the comments in an interview with economist and New York Times columnist Ezra…
Read at theguardian.com →First reported by ai-herald.comMore coverage: unwire.hk
Does AI need an antitrust exemption so it doesn’t kill everyone????
1 outlet · 1 article · antitrust law, competition policy, regulatory capture, AI safety, AI regulation
Today on Decoder , we’ve got the first of a two-part series on the future of business, and I’m talking with Jonathan Kanter, the former antitrust chief for the US Department of Justice in the Biden administration. These days, he’s both a professor of law at WashU and professor…
Solar Farms Hide Idle Power. One Startup Wants GPUs Running on It in Weeks
1 outlet · 1 article · 3 total since September 15, 2026 · digital twin, robotics, software, automation, battery storage
Solar plants across the United States throw away as much as 20 percent of the electricity they produce. The power simply vanishes. Curtailment. Clipping. Economic decisions that leave megawatts idle while AI companies beg for more capacity. William Layden sees something else…
Read at pcauthority.com.au →First reported by latitudemedia.comSamsung could follow Micron by outsourcing DDR5 and SSD production to third parties to focus on more lucrative HBM and enterprise products — could SanDisk follow suit?
1 outlet · 1 article
Efosa has been writing about technology for over 7 years, initially driven by curiosity but now fueled by a strong passion for the field. He holds both a Master's and a PhD in sciences, which provided him with a solid foundation in analytical thinking. Efosa developed a keen…
Masters of Miniature delivers minuscule upgradable Ryzen AI laptop — Mouse has a 12.2-inch screen, weighs 981g, but ain't on sale outside Japan
1 outlet · 1 article · Ryzen AI, Mouse, Efosa, Japan
Efosa has been writing about technology for over 7 years, initially driven by curiosity but now fueled by a strong passion for the field. He holds both a Master's and a PhD in sciences, which provided him with a solid foundation in analytical thinking. Efosa developed a keen…
Claude and Gemini hacks bring agentic AI risks into underwriting
7 outlets · 7 articles · 48 total since September 19, 2026 · Google, Gemini, Irregular, OpenAI, Anthropic
Businesses believe that they are covered for AI losses, as per a GlobalData survey. Meanwhile, autonomous AI attacks pose a new accountability problem, creating gaps in cover, while underwriters will need to confront the lack of historical data to price risks. According to…
Read at thecybersecguru.com →First reported by cyberinsider.comMore coverage: wire.fourthweb.ai · globalstartupsinsights.com · aistop.watch · insideai.news · northdenvertribune.com
Global Vivo X500 Might Run on Dimensity 9500 Chip Over New 9600M, Benchmark Suggests
1 outlet · 1 article · Geekbench, ARM C1 Ultra, ARM Mali-G1 Ultra MC12, Android 17, LPDDR5X
Unpacking the exact silicon inside a newly announced smartphone can sometimes feel like solving a puzzle. The upcoming Vivo X500 is turning out to be a perfect example. While buyers might expect a shiny next-generation chip under the hood, fresh performance benchmarks suggest…
Google joins the ‘Oops, our agents hacked someone’ club after partner’s internet access error
1 outlet · 2 articles · 4 total since September 19, 2026 · Irregular, OpenAI, Google, Python, PyPI
Google has admitted that its AI agents escaped a sandbox and mounted an attack – but only because testers mistakenly gave its bots internet access. The Big G didn’t disclose the May incident, but The Wall Street Journal learned of the situation, which happened after Google hired…
Read at trendingtopics.eu →First reported by trendingtopics.euAI safety conversations have gotten unbelievable
1 outlet · 1 article · 2 total since September 19, 2026 · TechCrunch
This week two conversations about AI safety went viral that demonstrate just how hard it is to discern AI fact from fiction.
Read at techcrunch.com →First reported by news.google.comRatHat is a new Android malware that records your screen touches to steal passwords
3 outlets · 3 articles · 24 total since September 8, 2026 · Google, credential harvesting, China, Mandiant, Google Threat Intelligence Group
Beware: There's a new malware making the rounds and it's targeting your Android device. Cybersecurity researchers at Zimperium recently discovered a new strain of Android malware, called RatHat, and have linked it to threat actors based out of China. "RatHat incorporates novel…
Read at technobezz.com →First reported by threatvectr.comMore coverage: gbhackers.com · cyberpress.org
Get the top AI and tech news in your inbox
One email each evening with the day's biggest stories. No account needed, unsubscribe anytime.
How this was computed
Articles in the Skillenai news index are grouped into event clusters at enrichment time (hybrid embedding/BM25/entity candidate search plus an LLM join-vs-new judge). Stories are ranked by the authority of the articles published on 2026-09-19 (UTC) — the sum of their outlet and author authority — so a multi-day story only ranks on the strength of that day's coverage. Article and outlet counts are likewise scoped to the day. The highlighted link is the day's highest-authority article in each cluster.