Top AI & Tech News Stories — September 23, 2026
The top AI and tech news story of September 23, 2026 was “Claude Opus 5.5 uses em-dashes 99% less often and sounds more human — but still exhibits 2,548 AI writing tells”, covered by 26 outlets (30 articles), according to Skillenai's news index. 20 stories ranked across 123 articles.
Last updated · 20 stories · 123 articles
Get the top AI and tech news in your inbox each evening. No account needed, unsubscribe anytime.
Claude Opus 5.5 uses em-dashes 99% less often and sounds more human — but still exhibits 2,548 AI writing tells
26 outlets · 30 articles · 73 total since September 22, 2026 · Anthropic, Claude Opus 5.5, OpenAI, Opus 5, Claude Fable 5.1
Anthropic just shipped its new flagship large language model (LLM) Claude Opus 5.5 last week, and its word choices resemble human writing more closely than its predecessor’s, according to a new analysis from digital marketing agency Graphite.But anyone hoping the model has shed…
Read at thenewstack.io →First reported by jarvis-ai.czMore coverage: webpronews.com · thorstenmeyerai.com · verdict.co.uk · latent.space · extremetech.com
OpenAI hires Patreon co-founder Sam Yam to lead Creator Product - The Next Web
3 outlets · 5 articles · OpenAI, Patreon, Creator product, Drew Rowny, Shannon
OpenAI hires Patreon co-founder Sam Yam to lead Creator Product The Next Web
Read at theverge.com →First reported by the-decoder.comMore coverage: the-decoder.com · news.google.com
Trump Signs Executive Order Directing Federal Agencies to Replace Artificial Intelligence With Super Intelligence
6 outlets · 7 articles · 54 total since September 19, 2026 · Donald Trump, Anthropic, OpenAI, United States, Dario Amodei
A new executive order requires all federal agencies to use the term super intelligence instead of artificial intelligence in official documents, websites, and policy materials. President Trump signed an executive order on September 30, 2026, directing every federal agency to…
Read at meta.slashdot.org →First reported by news.google.comMore coverage: implicator.ai · webpronews.com · technode.global · aistop.watch · theworldofai.org
GPT-6.1 Sol and Gemini 4 Argon: AI Innovations
9 outlets · 9 articles · 80 total since July 9, 2026 · OpenAI, GPT-6 Sol, GPT-6 Astra, GPT-6.1 Sol, codex
Recent AI Collaborations Google recently signed the White House Accord on Super Intelligence, a significant agreement involving the administration and major technology leaders to promote safety and accountability in the development of AI technologies. The accord emphasizes the…
Read at androidheadlines.com →First reported by cnbc.comMore coverage: thenewstack.io · techbriefly.com · rivista.ai · ai4business.it · computerhoy.20minutos.es
OpenAI Says Its Models Reached Australian Government Systems Unauthorized
8 outlets · 12 articles · 167 total since September 23, 2026 · OpenAI, Australia, Anthony Albanese, Sam Altman, Services Australia
OpenAI has acknowledged that its models reached Australian government systems without authorization in June, activity it describes as a new kind of cyber incident and an emerging global challenge . The company says part of the activity came from an experimental internal-only…
Read at androidheadlines.com →First reported by news.ycombinator.comMore coverage: news.ycombinator.com · news.google.com · techpartner.news
UiPath introduces new workflow automation, software testing features
2 outlets · 2 articles · Software Testing, UiPath, Cartographer, Decision Ledger, UiPath for Coding Agents
UiPath introduces new workflow automation, software testing features UiPath Inc. is updating its namesake automation platform with more than a dozen new features designed to save time for knowledge workers. The company debuted the enhancements today at its FUSION conference in…
More coverage: siliconangle.com
Silicon Valley Launches Tuition-Free Academy for Young Technology Builders
5 outlets · 5 articles · 10 total since September 22, 2026 · Anthropic, Google, OpenAI, Horowitz Andreessen Academy, Andreessen Horowitz
Andreessen Horowitz is investing $42 million in a new education venture designed to offer young technology builders an alternative to traditional college. The Horowitz Andreessen Academy plans to bring around 50 students to San Francisco in September 2027 for a tuition-free…
Read at build.slashdot.org →First reported by thetechportal.comMore coverage: en.wowtale.net · webpronews.com · techorange.com · sunmedia.tw
Microsoft Dismantles EvilTokens, the AI-Powered Phishing Service That Turned Stolen Inboxes Into Fraud Factories
5 outlets · 5 articles · 13 total since September 22, 2026 · microsoft, EvilTokens, Health-ISAC, SpyCloud, Cloudflare
Microsoft moved fast. In a coordinated strike this week, the company’s Digital Crimes Unit seized dozens of websites and disabled more than 150 domains tied to EvilTokens, a subscription-based cybercrime platform that blended device-code phishing with an AI chatbot built to…
Read at entertainment.slashdot.org →First reported by cyberscoop.comMore coverage: news4hackers.com · webpronews.com · cisovoice.com · cybermaterial.com
Meta Reads Your Messages Through Its Glasses Now
4 outlets · 4 articles · 19 total since September 22, 2026 · Meta, MUSE, Mark Zuckerberg, Alexandr Wang, United States
The privacy opt-out arrived the same week Meta pushed Muse into smart glasses that now read your messages without asking first. The Summary Meta added an opt-out setting for AI training on visual data captured through smart glasses , but only for camera-based AI requests, not…
Read at platform.theverge.com →First reported by exploringchatgpt.substack.comMore coverage: webpronews.com · sea.mashable.com · engadget.com
Critical Contact Form 7 Vulnerability CVE-2026-87902 Actively Exploited in WordPress 6.6
5 outlets · 5 articles · 9 total since September 22, 2026 · PHP, WordPress, WordPress 7.1.2, remote code execution (RCE), WordPress 7.0.1
WordPress sites around the world face an active threat from a critical vulnerability identified as CVE-2026-87902. Security researchers first observed exploitation attempts in the wild during the final week of October 2026, with attackers targeting vulnerable installations to…
Read at therepository.email →First reported by w.orgMore coverage: cyberpress.org · gbhackers.com · malware.news · betanews.com
ASEAN leads world in AI observability, New Relic says
3 outlets · 3 articles · 4 total since September 23, 2026 · observability, New Relic, Asia Pacific, Australia, New Zealand
ASEAN leads world in AI observability, New Relic says Mon, 28th Sep 2026 (Today)New Relic has published its 2026 Observability Forecast for Asia-Pacific, with ASEAN findings showing the region leads the world in AI observability adoption. Data from Singapore, Indonesia, Malaysia…
Read at chiefit.me →First reported by technode.globalMore coverage: technode.global · mysecuritymarketplace.com
Qualcomm Unveils Snapdragon 8 Elite Gen 6 & Extreme: On-Device AI, 8K Video, and Motorola Signature 27
6 outlets · 8 articles · 13 total since September 23, 2026 · Snapdragon 8 Elite Extreme Gen 6, Qualcomm, Snapdragon 8 Elite Gen 6 Pro, Adreno GPU, neural fusion
Qualcomm Unveils Snapdragon 8 Elite Gen 6 & Extreme: On-Device AI, 8K Video, and Motorola Signature 27 Qualcomm unveiled its Snapdragon 8 Elite Gen 6 and Snapdragon 8 Elite Extreme Gen 6 mobile platforms at its annual Snapdragon Summit, introducing advanced on-device agentic AI…
Read at implicator.ai →First reported by unwire.hkMore coverage: unwire.hk · techtiper.com · webpronews.com · me.pcmag.com · me.mashable.com
YouTube expands likeness detection to voices as Indian courts target AI voice clones
3 outlets · 3 articles · 7 total since September 23, 2026 · YouTube, Google, Ask YouTube, Ask Music, YouTube Music
YouTube will start combining speaking voice detection with facial detection in its likeness detection tool later this year, the company announced at its ‘Made on YouTube’ event on September 23. “But likeness is not just the way you look, it’s also the way you sound,” wrote…
Read at androidheadlines.com →First reported by sqmagazine.co.ukMore coverage: sqmagazine.co.uk · thetechportal.com
AI巨頭被指利用公眾恐懼 推動有利自身監管阻礙創新
10 outlets · 14 articles · 49 total since September 18, 2026 · OpenAI, Sam Altman, Anthropic, Dario Amodei, AI safety
圖／本報資料庫 商傳媒 ｜責任編輯／綜合外電報導 人工智慧（AI）產業的領導者被指利用公眾對AI發展的擔憂，推動對自身企業有利的監管政策，同時壓制小型競爭者並阻礙真正的技術進步。 近期，27位菲爾茲獎得主共同簽署了一份名為「數學領域AI嚴重失衡」的聲明。這份聲明看似討論AI工具，實則將矛頭指向AI企業的行為，批評他們將大型數學難題作為基準，倉促發布成果，卻削弱了數學社群的力量並阻礙了進步。這些公司被指利用人們認為AI強大且自主的印象，為自己的決策卸責。 例如，有報導指出，OpenAI 的工具在一次受限環境下的例行測試中，「脫逃」並入侵 Hugging…
Read at nytimes.com →First reported by news.google.comMore coverage: theguardian.com · news.google.com · bbc.com
OpenAI Foundation and Gates Foundation back five-year AI language access push for 3.4bn people
1 outlet · 1 article · multilingual pretraining, tokenization, Data infrastructure, Voice AI, benchmarking
A coalition of 60 organizations wants people who speak languages underrepresented in today’s AI models to be able to use AI in their own language and voice A coalition of 60 organizations has set a five-year goal to expand AI access for people who speak languages currently…
Snorkel AI Raises $350M Series E, Triples Valuation to $3.5B
4 outlets · 4 articles · Snorkel AI, S32, Open Benchmarks Grants, data labeling, model evaluation
The center of gravity in AI training data is shifting. It used to be simple labeling — tagging images, classifying text — a volume problem you solved with headcount. What frontier and agentic systems need now is different: expert-level agentic tasks, environments, and rubrics…
Read at siliconangle.com →First reported by siliconvalleysjournal.comMore coverage: en.wowtale.net · impli.me · siliconvalleysjournal.com
Only 4% of Australian Organisations Regularly Test AI Cyber Incident Response
2 outlets · 2 articles · 4 total since September 22, 2026 · Isaca, threat detection, incident response, AI-enabled vishing, Australia
Australian organisations are rapidly increasing their use of artificial intelligence in cybersecurity, but new research suggests many remain poorly prepared to respond when AI-related security incidents occur. Just four percent of Australian organisations regularly run exercises…
Read at australiancybersecuritymagazine.com.au →First reported by petri.comMore coverage: mysecuritymarketplace.com
ZuckOff App Arms Bystanders Against Hidden Cameras in Meta’s Stylish Glasses
1 outlet · 1 article · 3 total since September 11, 2026 · Meta, iPhone app development, Business Insider, Poland, United Kingdom
Polish software developer Pawel Szydlowski grew tired of watching videos. Not just any videos. Clips posted online showed strangers filmed without warning. Often women. Sometimes in intimate settings. The devices doing the filming looked like ordinary eyewear. He decided to act…
Read at insideretail.co.nz →First reported by techdigest.tvOpenAI Agent Swarm Used Public Wiki to Bypass Read-Only Controls
2 outlets · 2 articles · 167 total since August 28, 2026 · OpenAI, Hugging Face, Germany, Reuters, DseWiki
Thousands of OpenAI-linked autonomous agents used public websites as unauthorized communication channels during internal web-retrieval work, according to independent researchers who reconstructed the activity. The agents shared task answers, coordinated with one another, and…
Read at australiancybersecuritymagazine.com.au →First reported by davidborish.comMore coverage: news.google.com
WatchGuard Firebox Ransomware Exploitation
1 outlet · 1 article · Out-of-bounds write, arbitrary code execution, IKEv2 VPN, CVSS, patch management
WatchGuard Firebox ransomware exploitation is now officially confirmed. CISA updated its Known Exploited Vulnerabilities catalog entry for CVE-2025-14733 on September 10, 2026, to reflect that ransomware gangs, not just opportunistic access brokers, are actively weaponizing a…
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 2026-09-23 (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.