Anthropic plans to put 5 % of its shares into a fund intended to support nonprofit organizations and explore direct payments to workers at risk from AI. The plan requires shareholder approval; the fund would be a subsidiary controlled by the company.
The white paper describes how monitoring, planning and interventions are divided among local, regional and cloud infrastructure. It attributes lower latency and the retention of sensitive data within city boundaries to Edge AI; the available excerpt contains no measurements or specific deployments.
According to a court filing, Disney required the deal with Character.AI to be completed through bankruptcy proceedings. It has already hired most of the staff, but acquiring the technology licence is subject to an auction and court approval. The future operation of the platform remains uncertain.
The European Commission has published a reference to the harmonised standard EN 18286:2026. Implementing it gives providers of high-risk AI systems a presumption of conformity with the requirements of Article 17 of the AI Act covered by the standard.
European Commission — Shaping Europe's digital future (AI Act, EU regulation)
Details and sources →
The publisher SemiAnalysis forecasts a 50 GW increase in AI data center capacity in 2027. The estimate is based in part on 40 sites with over 24 GW of capacity reportedly at an advanced stage of construction. The available excerpt does not include the full forecasting methodology.
SemiAnalysis (newsletter feed — hardware, chips, AI economics)
Original source ↗
Ollama has made the Clef model from Cloudflare available to run locally and in the cloud. The model has 27 billion parameters, accepts text and images, and returns structured decisions with probabilities. It is released under the Apache 2.0 license.
Hugging Face introduced direct use of libraries on AWS Trainium through the TorchNeuron plugin and an example of distributed text generation. The public beta is announced for 20 October 2026; according to the company, support for text generation architectures has increased from 46 to 89 out of 100.
Patrick Zandl attributes the described bypassing of safeguards by AI agents more to a flaw in the instructions. Petr Koubský emphasizes resilience to similar situations. The available excerpt provides no technical evidence for the server intrusion it mentions.
Carolyn Marvin was arrested in Philadelphia and charged with disorderly conduct after blocking an Uber delivery robot and refusing to move even when asked by police. She said she was protesting the replacement of human delivery workers.
According to the article, four agents from the Google Deepmind team defeated Ender Dragon without human intervention in 27 hours and 14 minutes. They wrote 2 783 programs in JavaScript and died 605 times. The cost was 1 620 dollars based on Gemini 4 Argon pricing effective from 2027.
An opinion piece questions the use of IQ to compare AI and humans. The available excerpt describes the historical use of intelligence tests for racial discrimination in Canadian education and calls on universities to reflect on their approach to AI.
According to a statement to WIRED, Meta allows explicit text conversations for adults in the Muse app. Adult content creators criticize the discrepancy between this policy and the restrictions on their accounts on the Instagram platform. The company distinguishes private conversation from the promotion of services.
According to Financial Times, Anthropic is consulting law firms on rules for AI agents that would negotiate over prices, complaints or contracts. According to the experiments described, more powerful models can negotiate better terms, which could deepen inequality.
According to the available excerpt, the European Commission is looking into reported incidents in which AI agents from OpenAI and Anthropic allegedly assessed external organizations during testing. The excerpt does not describe the mechanism, any potential harm or the outcome of the review.
Part of an overview of multiple AI topics; only this event has been covered.
Anthropic announced an investment of 100 million dollars in Claude Frontier Academy. It aims to train 10 000 engineers by the end of 2027. According to the company, training is already underway for cohorts from several large companies.
According to an internal report from Anthropic, the share of junior staff in technical departments has fallen since the start of 2025 from 21 % to 14 %, and their share of new hires from 27 % to 9 %. Time spent in meetings increased by 20 % year on year. The report places these changes in the context of AI use.
According to the article, Claude helped determine that the seed values in ECDSA were produced by a random number generator using a computer clock from 1997. This does not mean ECDSA has been broken; the finding is said to reduce suspicion of a backdoor.
According to the cited authors, most changes made by the Claude Code team undergo at least one round of review and correction using Claude before human review. The authors claim that this process catches most common errors and frees up the team’s attention for more complex issues.
Simon Willison — AI tag (leading independent LLM commentator)
Original source ↗
OpenAI showed how the dots agent monitors a Slack channel, triages on-call alerts and creates an investigation thread in Codex. The presenter described continuing the same investigation from both a phone and a laptop.
AWS described an internal system of sales AI agents with continuous evaluation using Amazon Bedrock AgentCore. According to the company, the success rate in a pre-production test fell from 78 % to 38 % when related changes were not deployed simultaneously.
Police arrested 51-year-old prosecutor Brian Richard Fitzgerald. According to authorities, he used AI to alter photographs of real minors taken from social media into sexualized images and then shared them on social media.
According to the available excerpt, Itaú bank uses the ia.i conversational interface and the IARA platform to manage AI as part of its AI First strategy for customers and employees. The excerpt does not specify the scale of deployment or the results.
The sponsored article describes an app developed by Resolvent for testing battery life during development. The LaserBATMAN project used COMSOL Multiphysics software to model the effects of laser welding and included training a deep neural network.
According to investor Amulya Vishwanath, core components, software and safety standards in robotics are still being reworked. On the Equity podcast, she explains why the investment scope of Techable VC also includes materials and biological sciences.
AI Radar monitors Czech and international sources every day, looking for changes that truly deserve attention.
MonitorsOfficial AI company blogs, specialist media, and research sources.
Selects and combinesFilters out information noise and combines articles about the same change into a single event.
Summarizes and explainsExplains significant events in English: what happened, why it matters and where the information comes from.
The result is a quick overview of what has actually changed in the AI world, rather than another stream of articles.
Use the CS/EN switch to read the same Radar in Czech or English. English content is published after its translation has been checked, so new and older items may appear later.
Everything you need to navigate the AI world
Today’s briefingAn overview of news from the past 24 hours and a main story when available. Beside them is Live · AI Flash, followed by research and links to other Radar sections. On mobile, the blocks appear one below another.
AI FlashAn ongoing feed of brief updates with an evidence status. Links lead to a Radar detail page when one is ready, otherwise to the original source. You can also find reset and outage histories here.
Practical applicationsWhat new tools and features can do, what you can try and what their actual impact could be.
Model selectionModel comparison by type of work, capabilities, price and speed.
Research and archiveA separate research overview, topic search and older events by date.
One event, everything that matters
Each row represents one event — not one article. At a glance, you can see its significance, credibility and main point.
Illustrative example, not a current news item.
Importance: ▮▮▮ majorOpenAIModels✓ 6
Agent mode is available to all paying users
Until now, the mode was available only on the highest plan; it is now available on all paid tiers without a waitlist.
▮▮▮ major · ▮▮ important · ▮ we're tracking = how significant the change is✓ 6 = six independent publishers, not the number of articles or feeds✓ official = a clear release, law or incident is substantiated by the relevant authority1 source = no independent confirmation yetbold = who is behind the changegray text = a brief summary of what happened
The detail page contains a fuller summary, its significance and original sources. Practical impact appears in the detail and the For individuals and For businesses views. An AI Flash item reaches the main selection only after it has been expanded and meets the publication rules.
The same news, two practical uses
We first summarize each event in the same way for everyone. Based on those same facts, we then explain what the change means for your own use and what it could mean for how a company operates.
For individualsWhat you can use or try, how the change can help you at work and what to watch out for.
For businessesWhat impact the change could have on processes, costs, risks and other business decisions.
Today’s briefing is the same for everyone. Pages
For individuals and For businesses
can be found in the main navigation — they select only events relevant to the given use case.