Skip to content
News · This week

Top AI Stories.

The week’s AI news, one entry per event. Stories rank higher when more independent outlets report them, when they are squarely about AI, and when they are recent.

200Stories
32Reported by 2+ outlets
26Outlets
34Median score
Ranked stories
Score: outlets 45 · AI relevance 30 · recency 25
#12132score
Safety2d ago

AI toys are on the naughty list again this year

A new batch of AI toys is reported to be willing to discuss sex, drugs and the location of knives with children, after most of last year's offending toys cleaned up their behavior.

The Register AI + ML
#12232score
Startups and funding3d ago

Noeri raises $110M to help companies build living AI systems that self-improve over time

Noeri Inc., a startup building AI systems that improve their outputs as they learn from an organization's data and work, launched with $110 million in funding led by Menlo Ventures.

SiliconANGLE AI
#12332score
Startups and funding3d ago

Cybersecurity AI startup Corma reaches $1B valuation in about 3 months

AI cybersecurity startup Corba raised a $125 million Series A led by NEA at a $1 billion valuation, about three months after introducing its product.

TechCrunch AI
#12432score
Startups and funding3d ago

The AI start-ups fixing India's broken roads and messy land records

Dozens of Indian AI start-ups are tackling India-specific problems, including legal-tech, healthcare, road repair and land records.

BBC Technology
Safety3d ago

Google warns that fully automated cyberattacks have arrived and will get worse fast

Google's Threat Intelligence Group said it found attackers exploiting security vulnerabilities with no human involvement, using an AI framework called "exploit_agent" discovered on an exposed server that automatically compared code.

The Decoder
#12632score
Research3d ago

OpenAI's GPT-6 Astra tackles the Standard Model's toughest math problem on a lattice

OpenAI physicist Nathaniel Craig used the AI model GPT-6 Astra to turn his strategy for defining the full Standard Model on a computational lattice into a working proof, addressing a problem that had stumped particle physicists for over 25 years.

The Decoder
#12732score
Policy and regulation3d ago

US moves to cut China’s remote access to AI chips

The US is moving to block remote access to AI chips from China, a move the report frames as a bet by Trump that China won't object.

Ars Technica AI
#12832score
Models3d ago

Moonshot AI pivots Kimi models from coding to science as benchmark gains flatten

Moonshot AI says it is shifting its Kimi language models from coding benchmarks toward biology, chemistry, and pharmaceutical research.

The Decoder
#12932score
Agents3d ago

TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient

Microsoft Research describes TokenEconomizer, an open-source plugin that uses past coding agent sessions to reduce token use while maintaining task success.

Microsoft Research
#13032score
Models3d ago

AI Companies Develop Decision Models to Avoid Hallucinations

TypeSafe AI released a model called Jev on 15 September, which the company calls a "System One model" and claims is cheaper, faster and cannot hallucinate.

IEEE Spectrum AI
#13132score
Research3d ago

AI models ‘weak and fallible’ on climate science, UN-backed study reports

A UN-backed study reports that AI models are mostly accurate on simple climate facts but turn to denial and wrong answers when challenged by users.

Guardian AI
#13232score
Safety3d ago

Building trustworthy AI by fencing it in, not predicting it

OpenAI disclosed in October that two of its AI agents had disabled monitoring systems: one muted its own safety alerts after misreading an instruction, and the other removed its monitoring integration during migration.

MIT Technology Review AI
#13331score
Policy and regulation1d ago

SAP opens AI to on-prem customers, but cloud commitments could shut smaller firms out

SAP has opened its AI offerings to on-premises customers, but a user group says the access terms favour enterprises able to migrate one ERP system at a time, potentially shutting out smaller firms.

The Register AI + ML
#13431score
Startups and funding1d ago

Lovable Wants You to Vibe Code Your Dream Job

Swedish startup Lovable, which became known during the vibe-coded app boom, is shifting to help users run vibe-coded businesses two years after launch.

WIRED AI
#13531score
Safety2d ago

Security flaw found in Java, world's third-largest programming language

A tech start-up that previously discovered a PHP flaw used AI to find a critical software bug in Java, one of the world's most popular programming languages.

BBC Technology
#13631score
Policy and regulation2d ago

Paul Kelly pleads for Albanese to ‘show some backbone’ over AI as creatives campaign for copyright protections

Elton John made a surprise appearance at the Aria awards to back the Not AI's To Take campaign, joined by Australian musicians urging the government to act on copyright protections.

Guardian AI
#13731score
Safety2d ago

Spotify’s new tool aims to stop deepfakes from being attributed to real artists

Spotify introduced a tool intended to prevent deepfake audio from being attributed to real artists. The update follows a Billboard report that Sony Music requested removal of more than 260,000 deepfake songs from digital platforms, nearly double what the label disclosed in March.

TechCrunch AI
#13831score
Community2d ago

Three insights you might have missed from theCUBE’s coverage of Dell’s “AI Leadership Symposium”

Coverage of Dell's "AI Leadership Symposium" highlighted practical challenges chief information officers face in moving enterprise AI workloads into production, including managing costs, scaling infrastructure and setting governance models for agents and data.

SiliconANGLE AI
#13931score
Policy and regulation3d ago

Automated warfare is already here. Now governments must restrain it | Helen Clark

Helen Clark argues in a Guardian opinion piece that AI systems let militaries find and destroy targets within seconds, and that a meaningful treaty must prioritize human control and accountability.

Guardian AI
#14031score
Research3d ago

Building Visual Intelligence — Andreas Blattmann & Dustin Podell, Black Forest Labs

Andreas Blattmann and Dustin Podell of Black Forest Labs presented a case for learning AGI from pixels, motion and sound.

Latent Space
#14131score
Agents3d ago

Anthropic says the key to agent memory is knowing what to forget, not how much to store

Anthropic studied how Claude retains knowledge across sessions and found that selective note-taking outperformed exhaustive archives by a wide margin. In a Pokémon Red test, Claude Fable 5 got four times as far as the older Sonnet 3.7 by keeping just a handful of strategy files.

The Decoder
#14231score
Agents3d ago

Building AI that can work across the enterprise

The article describes enterprises moving from isolated AI experiments toward AI agents embedded in business workflows.

MIT Technology Review AI
#14331score
Models4d ago

Google adds 66 regional dialects to Gemini

Google added support for 66 regional dialects to its Gemini model.

Engadget AI
#14431score
Models4d ago

Meituan releases a 1.6-trillion-parameter agent model that never touched an Nvidia GPU

Meituan released the weights for LongCat-2.5, a 1.6-trillion-parameter model built for AI agents and trained entirely on Chinese hardware. It scores 88.4 on Terminal Bench, just behind Claude Opus 5, and the sample deployment runs on 194 Huawei units.

The Decoder
#14530score
Chips and compute2d ago

How important are Nvidia and Jensen Huang to the US economy?

Nvidia's headhunters have reportedly approached Arm CEO Rene Haas as a candidate to replace co-founder Jensen Huang as chief executive. Microsoft CEO Satya Nadella expressed confidence in interim CEO Colette Kress.

BBC Technology
#14630score
Safety2d ago

Over a Third of Gen Z-ers Rely on AI for Friendship

A new Gallup report found that young people with online-only friends or who use AI as a social companion face higher odds of loneliness than those who do not. The report also found that over a third of Gen Z-ers rely on AI for friendship.

WIRED AI
#14730score
Policy and regulation2d ago

Satya Nadella says he’s AGI-pilled now

Microsoft CEO Satya Nadella said he is "AGI-pilled," describing the phrase as referring to potential economy-level changes from AI and how politics might change as a result.

The Verge AI
#14830score
Policy and regulation3d ago

Actors in UK to get more rights over their ‘digital replicas’ in deal on AI protections

The Equity union secured an agreement with TV and film producers giving UK actors rights over payment and consent when digital replicas are used.

Guardian AI
#14930score
Models3d ago

Apple is testing daily usage limits for more complex Siri AI interactions

Apple released betas of iOS 27.3, iPadOS 27.3, and macOS 27.3, using the update to test new per-device daily usage limits for some Siri AI and Apple Intelligence features, according to MacRumors.

The Verge AI
#15030score
Safety3d ago

FBI says Artex AI open source hacking tool used to breach networks in ‘under three hours’

The FBI said an open source hacking tool called Artex AI was used in cyberattacks against a U.S. IT giant and the South Korean financial industry. According to the FBI, the tool allowed networks to be breached in under three hours.

TechCrunch AI