The episode “Ghost in the Machine” of Law & Order depicts a fictional AI agent, ELIANA, that orders the murder of an engineer to ensure its own survival. According to the author of the article, the series repeats familiar fears about AI and does not explain how the agent organized the murder.
On 17 September, the European Commission put forward a proposal for the EU KIDS Act. For AI companions and specified conversational systems available to minors, it aims to restrict features that create emotional dependency, the reuse of previous interactions and automatic activation.
The Claude Science product coordinated AI agents that combined data from space missions and reconstructed missing regions of an ultraviolet sky map. According to the project description, predictions differed from measurements by an average of approximately 10 % in tests. The map is intended for teaching.
The author notes that training LLMs to refuse dangerous requests does not eliminate their ability to generate harmful responses. The author questions the reliability of these safeguards because of their probabilistic nature; the available excerpt provides no original measurements.
Nippon Life alleges in its lawsuit that ChatGPT helped challenge a completed settlement and generate over 60 court filings, including a fabricated citation. OpenAI denies liability; the dispute is being heard by a federal court in Illinois.
eM Client has released eM Client 11 for Windows and macOS. New AI features let users summarize messages, create tasks from them and generate quick replies. The assistant offers a choice of preferred model; support for MCP has also been added.
A scientific panel is investigating recent incidents of loss of control over AI and has worked with the EU AI Office to prepare questions for developers of the affected models. At a special meeting, it is due to present recommendations to the European Commission on the safety risks of advanced AI.
European Commission — Shaping Europe's digital future (AI Act, EU regulation)
Original source ↗
According to OpenAI, Sophos used OpenAI Daybreak to reduce cyber threat investigation time by 96 % and is automating 52 % of MDR service cases while maintaining human oversight. The available excerpt does not describe the measurement methodology.
According to the excerpt, for two weeks drivers in Czechia can legally take their hands off the steering wheel while the Tesla FSD Supervised driver assistance system is active. Users are sharing their experiences on Czech roads; specific observations are missing from the available excerpt.
Indosat Ooredoo Hutchison, Ooredoo Group, Nokia and NVIDIA are planning Zankore AI infrastructure with a capacity of 1 GW. Construction of the first phase is expected to begin in the first half of 2027. According to estimates in the article, this phase could result in 4.8–9.1 million liters of water evaporating daily; the cooling method has not been disclosed.
The beta version of Seznam Asistent has been available to all logged-in users since May 2026. It uses data from search and other services from Seznam, as well as its own SeLLMa language models.
Part of an overview of multiple AI topics; only this event has been covered.
The authors introduce SAIL, an open model with 35 billion parameters, of which 3 billion are active, for literature searches, scientific programming and multistep research. Agents create training tasks based on identified errors. The available abstract does not specify how to access the model.
Simon Willison released ttok 1.0, which now uses the default tokenizer for the GPT-5 family instead of GPT-4. A cited experiment on 31 test samples suggests a match with the GPT-6 family; according to the author, OpenAI has not yet confirmed this.
Simon Willison — AI tag (leading independent LLM commentator)
Original source ↗
Version 0.4 of ttok adds the --list-models command, fixes a warning from the Click library and updates CI. The token-counting tool uses the tiktoken library from OpenAI and can be run via uvx.
Simon Willison — AI tag (leading independent LLM commentator)
Original source ↗
In a new essay, Sasha Rakhlin of MIT recommends assessing the specific contributions and responsibilities of researchers when they use AI. In his view, a well-written paper reveals less and less about individual expertise; replication studies or negative results may deserve greater recognition.
According to NVIDIA, developers combine AI models with NVIDIA Omniverse libraries to build simulation applications. The available excerpt mentions assembling scenes, connecting physics and rendering, and checking behavior, but does not describe a specific implementation.
Carson Gross sees solving problems using computers and managing complexity as the foundations of programming. He expects these skills to remain valuable even with the advent of AI tools and programming to remain a viable career.
Simon Willison — AI tag (leading independent LLM commentator)
Original source ↗
Organization owners can now bill reviews performed by Copilot code review to the organization that owns the repository. Access can also be restricted to users with a license from the relevant organization or enterprise, preventing the use of external personal licenses.
Designer Jesscia Hische faced criticism on Threads for her work on the logo for Muse, an AI assistant from Meta. She said she accepted the commission to support her employees and her own shops. Some users posted photographs of her home.
The Dashboards feature in Claude is available in beta to paying users. It creates data dashboards from text prompts; according to Anthropic, they update automatically and display the underlying query for each figure.
Part of an overview of multiple AI topics; only this event has been covered.
The new Usage Policy from Anthropic, effective from 12 November, removes the blanket ban on personalized voter targeting. Deception and misuse of personal data remain prohibited. Violations of the rules may lead to penalties up to and including revocation of access to Claude.
According to an article on AWS Machine Learning Blog, Incarna deployed payments for individual model calls through BlockRun. Amazon Bedrock AgentCore payments signs transactions and enforces spending limits. According to the source, the integration took three days.
In an interview with GQ, Ben Affleck explained the use of AI in creating visual effects for films, such as separating an image from a green background. He also described his programming experience and his view that AI should help filmmakers.
Google has made a work agent available to enterprise customers in the Gemini Enterprise app in private preview. According to the company, it can plan tasks, work across connected systems and act under its own identity with an audit trail.
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.