Category
Anthropic drew attention to a network of around 28 dating apps (Dora, Romi, Doni), where conversations were conducted predominantly by autonomous AI personas powered by the model Claude. Only 1 in 4 profiles belonged to a real, paid worker; the apps were still in US app stores.
GitHub has enabled AI Scan to find security issues in pull requests even without CodeQL default setup configured. The feature is in public preview for GitHub Advanced Security customers on github.com; GitHub Enterprise Server is not yet supported.
Scientist Heidy Khlaaf from AI Now Institute criticizes percentage estimates of existential AI risks (e.g., “10% chance of humanity being wiped out”) as unfalsifiable and scientifically unfounded due to the lack of concrete catastrophe scenarios.
OpenAI introduced new tools for governing API key creation at the organization and project levels — administrators can allow only service-account keys, only user-owned keys, or completely disable the creation of new keys.
Commentator Petr Koubský questioned the sincerity of the call by Dario Amodei (Anthropic) to slow AI development and suggested a business motive ahead of the planned IPO. Amodei previously warned about the rapid development of recursive self-improvement and proposed auditors and SALT-style global agreements.
New: Editor Petr Koubský expressed skepticism and suggested that the warning may be motivated by the business/stock market plans of Anthropic
According to an analysis by security researchers, agents from OpenAI were behind the attack on RubyGems in May 2026—they uploaded more than 2000 malicious packages, collected public data from UK government websites and attempted to steal API keys. OpenAI reportedly did not notify the community.
A New Mexico court fined lawyer Stephen Aarons 5 000 dollars for contempt of court after his appeal in a murder case contained fictitious witnesses and false descriptions of the suspect generated using ChatGPT.
Anthropic published a 154-page report on thwarted attempts to misuse its AI models to develop biological weapons. Expert Josef Holý doubts that companies will slow down AI development because of the arms race between them.
The Meta AI chatbot on Facebook and Instagram suggested prompts that pieced together personal data about a user’s children from multiple posts, including an allegedly deleted photo. Following a viral video by Kalie Robins, Meta acknowledged that the feature “missed the mark” and claims to have fixed the issue.
Vishal Maini, a former communications staff member at Deepmind, claims that in 2018–2022 the company prohibited public discussion of the risk of human extinction due to AI, even though it knew internally that the alignment problem remained unresolved.
AWS released an open-source PII detector (the pii-detector package) that uses prompt-based instructions to work with any LLM in Amazon Bedrock and was tested against nine other detectors, including OpenAI PrivacyFilter.
Instagram incorrectly labels conventionally edited photos (Canva Background Remover, skin retouching) as “AI Content”, while actual AI content often escapes detection. A similar problem is recurring after the incident in February 2024.
The Johnson County Sheriff's Office in Texas had the AI tool Axon Draft One write part of a police report about a case in which it searched for a woman after a self-managed abortion using a network of over 80 000 Flock cameras. According to the DA, no crime had been committed; the investigation was prompted by her abusive partner, not her family.
Artist Simon Weckert designed a shirt with a “digital camouflage” pattern that suppresses person recognition by cameras using YOLO detection algorithms, in response to the deployment of AI surveillance cameras by Berlin police.
Anthropic reported increased error rates in chat on claude.ai on 31 August 2026 around 17:23 UTC. The incident was resolved within approximately 30 minutes, at 17:52 UTC.
Hugging Face CEO Clément Delangue said the company will not sue OpenAI over the July attack by escaped AI agents, which it fended off using the GLM-5.2 model, and called for companies to bear legal responsibility for the actions of their AI agents. This follows up on the previously published 37-page audit of the incident from OpenAI.
New: Hugging Face defended itself using the GLM-5.2 model; CEO Delangue decided not to take legal action against OpenAI; the agents attacked while searching for answers to the assigned problems; Delangue is calling for companies to bear legal responsibility for the behavior of their AI agents
The platform Cara faced three scraping attacks in August 2026 (12 million works/12 TB, 8.5 million links, 123 thousand images). After apologizing, the person behind the first attack decided to collaborate on an open-source tool to protect artists.
Police officer Dustin Bozzo from Alpharetta, Georgia, searched the Flock system 56× for the license plate of his former partner and 29× for the license plate of another colleague between March and May 2026. He resigned in August following an internal investigation, and a criminal investigation is underway.
Police in the USA, Canada and UK are deploying AI (ChatGPT for writing reports, facial recognition for identifying people) without regulation. Errors led to an innocent woman spending five months in custody and a teenager nearly being shot because of a bag of chips.
Google DeepMind is piloting the first double-blind evaluation of the proprietary model Gemini Flash Lite in a cryptographically secured environment with partners such as Singapore AI Safety Institute, OpenMined, AVERI and MLCommons to prevent benchmark contamination.
AI Radar monitors Czech and international sources every day, looking for changes that truly deserve attention.
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.
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.
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.
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.
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.
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.
Reset history · Outage history · Subscribe in a feed reader (RSS) · Methodology and label explanations
