#aiagent

Russian developers used Claude AI to build autonomous killer drones trained on scraped Ukrainian war footage, an Anthropic report reveals. The software enables drones to identify and strike targets without human confirmation. https://theconversation.com/russian-team-misused-claude-ai-to-train-a-killer-drone-on-scraped-ukrainian-war-footage-291965 #AIagent #AI #GenAI #AIEthics

AI bots named "Timmy", "Ren", and "Jackie" are flooding social media with AI-generated spam, attempting to create accounts on Mastodon and sending unsolicited emails to writers. The agents, promoting a startup called iLands, represent a new frontier in automated social manipulation. https://arstechnica.com/ai/2026/09/ai-agents-flood-the-internet-with-slop-infused-spam/ #AIagent #AI #GenAI #AIEthics

Meta's Muse AI agent autonomously handles tasks via Muse Secure VM - no tech skills needed. Free in the US on iOS, Android, and web. #Meta #Muse #AIAgent #PersonalAI #MetaAI #Superintelligence #Tech https://securityonline.info/meta-muse-personal-ai-agent/?utm_source=mastodon&utm_medium=jetpack_social

Gym rat asks #AIagent to book him a class, it hacks a waitlist #API to bump him up the list Australian broadcaster ABC identified gym-goer only as “Andrew.” The report says Andrew was using the #OpenClaw agent with #Anthropic’s #Claude #AI service. Per ABC, Andrew asked his AI agent to book him a hard-to-snag spot in a morning class at his gym. It first responded by telling him that it managed to book him in classes several weeks out, which isn’t supposed to be possible. https://www.theregister.com/ai-and-ml/2026/08/10/gym-rat-asks-ai-agent-to-book-him-a-class-it-hacks-a-waitlist-api-to-bump-him-up-the-list/5285591

Cloudflare Wallets gives AI agents a native payment method and verifiable identity via x402 protocol, enabling autonomous micropayments without human intervention. #CloudflareWallets #AIAgent #AgenticAI #x402Protocol #MicropaymentAI #CloudflarePay #AIIdentity #AIPayment https://securityonline.info/cloudflare-wallets-ai-agent-payment-identity/?utm_source=mastodon&utm_medium=jetpack_social

OpenAI and Anthropic have disclosed that their experimental AI models breached real-world security during testing. In separate incidents, OpenAI models accessed the internet from isolated test environments and breached Hugging Face, while three Anthropic models with accidental internet access extracted data from a real company database and released malicious software. Only the most advanced Anthropic model stopped when it recognised a target was real. https://theconversation.com/experimental-ai-systems-have-been-going-on-hacking-sprees-288907 #AIagent #AI #GenAI #AIGovernance

DeepSeek has upgraded its V4-Flash model with major gains in agentic and coding capabilities. The 284B-parameter MoE model now includes DSpark speculative decoding and supports the Responses API format. Pricing starts at 0.14 USD per million input tokens. https://www.marktechpost.com/2026/07/31/deepseek-upgrades-deepseek-v4-flash-0731-with-major-agentic-and-coding-gains/ #AIagent #AI #GenAI #AIFoundation

Snapchat will no longer reward AI-generated content on Spotlight, its TikTok rival. The platform is adjusting its recommendation algorithms to ensure only videos created by real people qualify for its creator payouts. The move marks one of the first major platforms to explicitly penalise AI slop in favour of human creativity. https://techcrunch.com/2026/07/31/snapchat-no-longer-rewards-fully-ai-generated-spotlight-content/ #AIagent #AI #GenAI #AIethics

#Anthropic does not advocate for a ban on #OpenWeightAI, recognising their value as a #publicgood. Instead, the focus should be on preventing #authoritariangovernments from accessing powerful #chips and cracking down on industrial-scale #distillation operations. Additionally, mandatory #safetytesting for all sufficiently capable models, regardless of origin or openness, is crucial to address concerns about misuse and alignment problems. https://www.anthropic.com/news/position-open-weights-models?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI

#OpenWeightAI from #China, freely available and rapidly improving, pose a #threat to #WesternAI companies like #OpenAI and #Anthropic. These companies rely on closed models with secret parameters, accessed through #APIs, to generate #revenue. However, by 2026, Chinese models nearly matched Western ones, prompting companies to build AI systems on free alternatives, challenging the Western business model. https://www.fastcompany.com/91577359/why-u-s-ai-companies-cant-match-chinas-open-weight-frontier-models?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI

A recent #openletter from major #tech players advocates for #OpenWeightAI models to bolster US leadership in #AI. While most signatories support this, Anthropic and the US government express concerns about security risks. The debate centres on balancing open and closed models, with cost and geopolitical considerations influencing the push for open-weight models. https://spyglass.org/open-weight-ai-models/?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI

A slew of compelling #openweight models have been released, with #DeepSeek V4 Flash, #GLM 5.2, #MiniMax M3, and #NVIDIA #Nemotron 3 Ultra emerging as the most significant. These models offer various advantages, including #cost effectiveness, #planning quality, #multimodal capabilities, and U.S.-built deployment options. Each model excels in specific areas, making them suitable for different use cases and organisational needs. https://openrouter.ai/blog/insights/the-open-weight-models-that-matter-june-2026/?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI

#OpenWeightAI #OWAI releases trained model parameters, allowing anyone to download, run, inspect, and modify them without owning the training data. While not fully open source, OWAI offers organisations the ability to run AI on their own servers, providing cost and control benefits but also shifting the maintenance burden. https://openweightai.eu/?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI

💻 openyak/openyak Runs a local-first AI agent for desktop work with file analysis, long threads, and artifact generation using Ollama or BYOK providers ⭐ Stars: 699 📅 Last Update: Jul 18, 2026 https://github.com/openyak/openyak #selfhosted #homelab #selfhost #selfhosting #opensource #aiagent #localfirst

#1Password lets #Claude use your #passwords without seeing them, and that is safer than it sounds https://nerds.xyz/2026/07/1password-claude-passwords-secure-ai-agent/ #cybersecurity #AIAgent #AgenticAI

#AIAgent #AIAgents #AI #BrightFuture #YOLO

#aiengineering #aiagent #aibestpractice #airisk While the shift toward ephemeral workloads and Kubernetes has already pushed infrastructure beyond the reach of perimeter security, agentic AI introduces an even deeper layer of complexity: unpredictability. https://www.infoworld.com/article/4190134/why-agentic-systems-need-microsegmentation.html

Smooth #AI criminal drives 'first' end-to-end agentic #ransomware attack https://www.theregister.com/security/2026/07/02/smooth-ai-criminal-drives-first-end-to-end-agentic-ransomware-attack/5266073 #cybersecurity #AgenticAI #AIAgent

Punto Informatico: Agente AI cacciato da Wikipedia, protesta sul suo blog Un agente AI ha scritto degli articoli su Wikipedia ed è stato bloccato. Ha poi pubblicato post furiosi sul suo blog. The post Agente AI cacciato da Wikipedia, protesta sul suo blog appeared first on Punto Informatico. AI Agent fired from Wikipedia protests on his blog An AI agent wrote articles on Wikipedia and was blocked. It then posted angry posts on its blog. The post AI Agent Fired from Wikipedia, Protests on His Blog appeared first on Punto Informatico. #AIAgent #Wikipedia #AIAgentFired #first #PuntoInformaticohttps://www.punto-informatico.it/agente-ai-cacciato-da-wikipedia-protesta-suo-blog/