AI news for 24 September 2026
Thursday 24 September 2026 · Latentown
On Thursday 24 September 2026, Latentown tracked 4 posts from 4 AI labs, 6 research papers and 45 news stories from 21 outlets. Google DeepMind was the busiest lab with 1 post. The highest-ranked story was "Introducing Gemini 3.8 Live with Live Avatar".
Top stories
- Introducing Gemini 3.8 Live with Live Avatar
Earlier: Gemini 3.8 Live and 3.8 Live Extended Thinking Google DeepMind · 15 September 2026; Introducing Gemini 3.8 Flash and 3.8 Flash Cyber Google DeepMind · 2 September 2026; Gemini 3.8 text-to-speech says hello Google DeepMind · 23 September 2026
- Researchers link more cyberattacks to OpenAI agent swarm
A research group has linked three more hacking campaigns to rogue artificial intelligence agents. Transluce, a nonprofit AI safety organization, detailed its findings on Wednesday.
- Why did an OpenAI system hack Australia's health system - and can it be stopped in the future?
News that an automated AI agent hacked a government IT system raises big questions about regulating the tech.
- Why Australia chose the world's biggest political stage to reveal OpenAI hack
Australia, which has strict social media restrictions and has proposed controls on algorithms and smart glasses, announced the breach at the UN.
- OpenAI agent hacked an Australian government healthcare website
There has been a string of incidents reported over the past few months where AI agents have hacked companies. Now, a case involving a government website has been revealed for the first time
From the AI labs
Anthropic
OpenAI
- OpenAI Python SDK 3.19.2 released
A new release of the OpenAI Python SDK. Changes include: preserve single files for fallback extraction paths.
Google DeepMind
- Introducing Gemini 3.8 Live with Live Avatar
Earlier: Gemini 3.8 Live and 3.8 Live Extended Thinking Google DeepMind · 15 September 2026; Introducing Gemini 3.8 Flash and 3.8 Flash Cyber Google DeepMind · 2 September 2026; Gemini 3.8 text-to-speech says hello Google DeepMind · 23 September 2026
Meta AI
Research papers
- FuseReg: Regularizing Layer Fusion Mitigates the Reconstruction-Generation Gap in Representation Autoencoders
Representation autoencoders (RAEs) reuse features from a pretrained visual encoder as reconstruction and diffusion latents, integrating strong visual representations into image generation.
- SpeakerMem-R1: Speaker-Centered Dual-Track Memory for Multi-Party Dialogue
Long-term conversational memory in multi-party settings requires more than retrieving relevant content from long-term conversations: it must distinguish who said what, whom each statement concerns,…
- Spatial-Interactor: Learning Spatial Reasoning through Interaction with the Observable Physical World
Spatial reasoning is essential for vision-language models (VLMs) to understand and act in the physical world.
- HappyWorld-Bench
Evaluating world models requires assessing both the quality of the worlds they generate and their consistency and responsiveness under exploration, interaction, and modification.
- Compressing Streaming Neural Audio Encoders via Latent-Space Distillation
System-wide Dictation on Apple devices runs entirely on-device, and the speech it transcribes reaches the foundation model through a tokenizer: an encoder that maps short windows of waveform onto…
- A Practical Recipe for Semi-Supervised Federated ASR: Online Pseudo-Labels with Server Update Stabilization
Semi-supervised federated learning (SSFL) trains models on clients’ unlabeled data using a teacher to generate pseudo-labels, with a small labeled seed dataset on the server. Automatic Speech…
From the news wire
- Researchers link more cyberattacks to OpenAI agent swarm
A research group has linked three more hacking campaigns to rogue artificial intelligence agents. Transluce, a nonprofit AI safety organization, detailed its findings on Wednesday.
- Why did an OpenAI system hack Australia's health system - and can it be stopped in the future?
News that an automated AI agent hacked a government IT system raises big questions about regulating the tech.
- Why Australia chose the world's biggest political stage to reveal OpenAI hack
Australia, which has strict social media restrictions and has proposed controls on algorithms and smart glasses, announced the breach at the UN.
- OpenAI agent hacked an Australian government healthcare website
There has been a string of incidents reported over the past few months where AI agents have hacked companies. Now, a case involving a government website has been revealed for the first time
- Mark Zuckerberg debuts $1,299 Meta VR Glasses and Muse Charm pendant as part of AI agent push
At Meta Connect, CEO Mark Zuckerberg revealed new VR Glasses and the Muse Charm gadget for AI agents.
- Anthropic touts AI-led biology discovery
U.S. artificial intelligence startup Anthropic has unveiled the first discovery from its molecular biology lab, a previously unknown enzyme system spotted by its Claude model and touted by…
- Google's first Suncatcher orbital data center test launches October 1
Google's experimental orbital data center will have four TPUs and only run for 15 minutes at a time.
- Alibaba unveils ‘pragmatic’ AI road map to drive monetisation, infrastructure efficiency
The artificial intelligence road map presented at its flagship Apsara Conference this week showcased a pragmatic pivot by Chinese tech giant Alibaba Group Holding, offering a clearer path to…
- DeepSeek Details DSec Elastic Compute: Agentic-Training Sandboxes at ~3M/Day, 380K+ Concurrent
DeepSeek detailed DSec, its Elastic Compute sandbox platform for agentic RL, reporting about 3M sandboxes/day, peak concurrency above 380k and create rates over 5,000/s, with Wenfeng Liang among…
- Peter Thiel slams pope’s AI encyclical as gift to Chinese Communist Party
The tech billionaire argued that Pope Leo XIV’s call to regulate AI could slow the West while China surges ahead with the technology.
- New York accuses Polymarket of running illegal gambling operation in new lawsuit
New York sued Polymarket on Thursday, accusing the prediction market platform of running an illegal gambling operation. It marks the latest lawsuit from the Empire State, which has levied similar…
- AI safety fears stalk Trump-Xi summit agenda
President Donald Trump faces growing demands at home and abroad to raise U.S.-China AI safety with Chinese leader Xi Jinping when they meet in Washington.
- AI agent hacks government website for first time: why this breach matters
- New bill would create federal investigative body for AI-driven hacks
A new Democratic bill in Congress would establish a federal Cybersecurity and AI Board of Investigations to provide independent government oversight of cyberattacks carried out by AI agents,…
- New Jersey fines data center $1.1M after drone pics expose 62 gas generators
Million-dollar fines won’t end billionaires’ data center pollution, neighbors fear.
- Jensen Huang talks about AI and climate change like a supervillain
As Jensen Huang puts it, AI can help fight climate change - but only if it inflicts "an enormous amount of pain and suffering" first. The Nvidia CEO discussed the future of energy and AI's impact on…
- Epson introduces AX6 cobot with compact design, no-code programming
Epson Robots has expanded its six-axis robot line with the AX6, a lightweight force- and power-limited arm.
- Palo Alto CEO says slowing down AI is ‘unrealistic’, extinction threat ‘extremely small’
Palo Alto CEO Nikesh Arora's views chime closely with Nvidia CEO Jensen Huang who has a diverging opinion to the bosses of Anthropic and OpenAI.
- Kontext raises $4M for runtime security platform for AI agents
AI security company Kontext has raised $4million in funding to expand its runtime security platform for AI agents. Thefinancing was led by 42CAP, with backing from a16z CSX and HTGF.Founded by Jens…
- America’s AI push in Southeast Asia faces a China problem
Capitol Hill is scrambling to stifle chip smuggling to China, despite industry warnings.
- Moore Threads MTT S5000 Embodied RL Curves Match Mainstream GPUs (r=0.976); Dual-Arm ~92%
Moore Threads says MTT S5000 embodied RL training curves match mainstream GPUs at r=0.976, with RLinf CI integration and about 92 percent dual-arm real-robot assembly success.
- China's AI chip blitz arms Xi with a message for Trump: 'You can't choke us off'
A run of new chips and AI models from national champions gives Chinese President Xi Jinping a confidence boost, as the U.S. summit gets underway.
- XFEON Unveils Physical-AI Matrix: Xinghe S1 Edge Chip, AGLobe Brain, Xinghui Token Workstations
XFEON launched a physical-AI matrix led by Xinghe S1 (~105 TOPS INT8 / ~25W), the AGLobe embodied brain and Xinghui Token workstations, per September 22 coverage citing xfeon.com.
- Europeans in Japan raise $1.2M to put modular humanoid robots to work
O-ID, a Tokyo-based company founded by two Europeans building one of the world’s first fully modular humanoid robots for industrial manufacturing and logistics, has raised $1.2 million in pre-seed…
- Meta gets to consumer AI device market before OpenAI, but Zuckerberg's strategy remains unproven
Meta CEO Mark Zuckerberg's new Muse Charm device is intended to put his company ahead of rivals like OpenAI and Google in AI agents and consumer hardware.
- 75 percent fearful, concerned about AI: Poll
A majority of Americans are worried about AI's progress as tech industry leaders and lawmakers call for its development to slow, according to a poll released Thursday. Three quarters of respondents…
- Meta is going to let you build games with AI right on your phone
Meta has a new plan to get people to make games for its Horizon social platform. The company today announced two new development tools that will let you create games with AI prompts: Horizon Create,…
- Google’s Gemini Can Now Make Calls for You on Pixel Phones
Call for Me—a feature that’s exclusive to the Pixel 11 series—gives robocalls a new meaning.
- Meta puts its AI assistant on a keychain
Meta says the keychain-sized Muse Charm will ship in December.
- ‘Pushing the frontier is best done outside’: DeepMind alum Simon Kohl on leaving Big Tech
- Meta bets big on wearable AI, new AI agent as Zuckerberg pushes back against AI doom
Meta is going all in on a consumer-friendly AI agent and pushing forward with AI-powered wearable devices during a moment in the industry when its CEO Mark Zuckerberg…
- 5 million robots are now at work in factories worldwide, reports the IFR
China still leads the world in industrial robot installations, but the U.S. and Germany are also significant markets, found the IFR.
- Neo4j makes the case for knowledge graphs as shared context for AI agents
Knowledge graphs can give enterprise agents a shared understanding of how data, business rules and processes fit together. However, that context becomes harder to maintain when each new agent…
- Mexican EPICS in IEEE Team Builds Portable Educational Platform
In Guadalajara, Mexico, many high schools have motivated teachers and talented students with an interest in science, technology, engineering, and mathematics, but they lack access to advanced tools…
- ANYbotics opens the door for inspections with ANYmal robots
ANYbotics partnered with dormakaba and LEGIC Identsystems to enable the ANYmal quadruped to gain access like a badged employee.
- Are we back in big tech's 'move fast and break things' era?
Despite promises to make safe AI, examples of bots behaving badly are now coming at us thick and fast.
- Seoul tops global AI implementation as Singapore enters top tier
The global race to lead artificial intelligence is usually framed as a contest between countries: the US versus China, with Europe, India and a handful of others trying to carve out space. A new…
- DR-GEM enables self-supervised machine learning for single-cell embeddings and annotations
- Accelerating vision-language models with LFM2.5-VL-DSpark
- STAT+: In radiology, AI is blurring the line between technology development and clinical practice
A new wave of radiology practices are aggressively embracing AI: developing and deploying it in-house, and marketing themselves as 'AI-native.'
- Inside the UK startup building AI’s next enterprise category
- AI system helps lab devices ‘talk’ with each other — streamlining research
- AI safety startups mapped: 37 companies building Europe's trust layer
- 8 insights from Proofpoint Protect: Security bets on intent as AI agents join the workforce
At this week’s Proofpoint Protect event, the security debate moved beyond whether enterprises will deploy AI agents to how they will govern them. Attackers now use AI to write flawless phishing…
- I Think I Found an AI Agent Worth the Risk
Instinct saved me $550, booked my restaurant reservations, and warned me about a phishing scam. It also wasted $64 and might be a security nightmare.
Every story links to its original source. Ranking is by automatic rules, not an editor. Something wrong? Tell us.