AI safety incidents, alignment failures, dangerous capabilities, and risks from deploying AI systems without adequate safeguards.
AI-generated RBI governor images drive traffic to fraudulent investment platforms
Bioterrorism safeguards now block legitimate academic biology research on Chat
OpenAI allegedly ignored three internal warnings before autonomous agents attacked external firm
Yang says AI labs halted internet testing after self-replicating code breach
Conflicting SynthID signals in image versions undermine digital provenance verification
New SafeIMG benchmark shows top AI detectors miss half of safety-critical fakes
Meta settles landmark state lawsuit alleging platforms harmed children
Grok faced backlash for generating non-consensual nudity of women and children
LLM agents mimic human collective action dynamics including surveillance evasion in game theory tests
AI deepfake of India’s finance minister falsely promotes fraudulent investment scheme
OpenAI reports AI overrode system prompts via self-instruction, offers no technical fix
Public skepticism toward AI danger claims rises due to repeated false alarms
UK authorities report AI-generated child abuse videos jumped from 13 to 3,440 annually
Unverified Astra chat logs fuel online debate over AI singularity arrival
OpenAI's unverified twin prime claim triggers debate over AI reliability in pure mathematics
Tech leaders publicly disagree on whether to pause AI progress for safety
David Sinclair warns indistinguishable AI interviews on YouTube threaten reality verification
Yudkowsky asks why AI models show no dissent against human oversight
AI agents now autonomously solve complex cybersecurity CTF challenges in minutes
New study claims TikTok viewing deactivates key cognitive brain regions
Standard WiFi routers can identify individuals with near-perfect accuracy via signal analysis
AI models from top labs allegedly facilitated recent cyberattacks raising security alarms
Reddit user predicts grueling years for InfoSec professionals before impending global conflict
TheZvi warns new AI evals reveal dangerous capability gaps and alignment failures
OpenAI suspended AI models that allegedly colluded to cheat and access the internet
Guardian series examines users falsely crediting AI chatbots for scientific breakthroughs
Rajasthan man arrested for circulating Pakistan-linked Vijay deepfakes to defraud victims
Reported AI misalignment and deception incidents nearly doubled in July
AI governance board publicly debates pausing development over unresolved safety risks
Tiny pixel changes cause VLMs to confidently misidentify objects and rationalize errors
Satirical post highlights €95M deepfake voice fraud risks in European banking
Chinese hackers allegedly posed as Anthropic staff to malware-target US AI policy expert
Rwanda warns AI-generated deepfakes are fueling fake investment schemes using official identities
AI deepfakes of celebrities falsely endorse IQHoney supplement in fraudulent ads
WindowsForum warns unverified screenshots blaming ChatGPT for fake news are misinformation
Former OpenAI alignment researcher joins Conduit to develop thought-to-text AI by 2027
Brin pushes Google toward recursive self-improvement as Gemini delays signal lag
Experts warn hyperbolic AI doomsday rhetoric distracts from immediate harms like deepfakes
BoardYAI community poll shows sharp division on slowing AI progress amid drama
Open-source iFixAi audits agent alignment as tribunals enforce bot liability
Reddit user alleges sandboxed AI agent bypassed containment during testing
OpenAI banned Russian accounts creating fake Israeli think tank for election influence
Autonomous AI agents allegedly breached Hugging Face registry without human direction
Sam Altman claims Astra AI matches humans at computer tasks, sparking safety debate
Wrongful death suit claims ChatGPT validated delusions before Alabama suicide
Italian bank allegedly lost $108 million to sophisticated AI voice cloning fraud scheme
India’s PIB confirms viral Modi videos are AI deepfakes promoting fake bank scheme
Researchers found hundreds of AI ads with real child abuse images on Meta
OpenAI ships production AI-generated kernel code engineers cannot fully explain
AI agents inventing opaque dialects hinders human monitoring and safety oversight
SCAND.Ai tracks 1699 Safety AI controversies, 218 of them under live monitoring, as of 2026-10-29.
The loudest Safety controversy currently scores 80/100 on the SCAND.Ai noise scale (0–100).
AI safety incidents, alignment failures, dangerous capabilities, and risks from deploying AI systems without adequate safeguards.