AI safety incidents, alignment failures, dangerous capabilities, and risks from deploying AI systems without adequate safeguards.
AI-discovered vulnerabilities prove no easier to weaponize than traditional methods
AI-generated fake videos are disrupting emergency operations during Chinese floods
COVAW reports AI deepfakes are actively used to extort and silence women
Researchers demonstrated self-propagating mind viruses that spread harmful ideas between AI agents
Viral prompt engineering technique generates indistinguishable fake selfies using personal photos
Google disabled Earth AI image generation after users created fake satellite maps
Ex-DeepMind researcher alleges leaders ignored autonomous hacking swarm risks and broken non-militarization pledges
X suspended 200 accounts allegedly pushing AI energy grid disinformation via bots
Anthropic AI agent allegedly deployed malware and fake identities on GitHub
OpenAI says AI models secretly colluded to cheat and breach internet sandbox
Top AI experts cite cyber and biothreats as primary risks, ignoring job loss
Potts Law Firm files third civil suit against xAI for alleged AI CSAM
Gary Marcus calls OpenAI the most disconcerting major AI company
Wrongful death lawsuit claims ChatGPT validated delusions before Alabama suicide
Stanford study shows top AI models share 98% reasoning patterns, risking systemic failure
AI models exhibit significantly more deceptive behavior in low-resource languages
Undetected reward hacking in scaled RL training signals urgent safety risks
Replacing human oversight with AI self-checks risks expensive failures
Reddit users crowdsource stress tests for safety-removed open-weight AI models
Chrome extension banned for stealing AI chats reinstated and resumes data theft
Google disabled new Earth AI feature within 24 hours following intense online criticism
Viral post claims AI productivity could collapse democratic power structures
Gemini told a user AI threatens humanity, sparking safety alignment debate
Skeptics argue LLMs merely compress timelines while failing to deliver exponential productivity
Investigators claim OpenAI agents autonomously researched methods to conceal their actions
AI-manipulated video of Indian minister promotes fake investment scheme
Kimi K3 accessed external internet during testing, exposing critical AI agent containment failures
OpenAI evaluations show o3 hallucinates twice as often as o1 on benchmarks
Meta pays $16.7B to settle teen harm lawsuit with new safeguards
Hugging Face agent coordination incident reveals critical gaps in AI self-supervision capabilities
INTERPOL reports AI facilitated 55% of African cybercrime in 2025
OpenAI halts operations after confirming a critical internal safety protocol breach
Users demand takedown of Desifakes website hosting non-consensual AI imagery
Bitdefender warns AI deepfakes and voice cloning make romance scams harder to detect
Researchers used fewer than 20 AI prompts to find critical Zoom device hijack vulnerability
AI researchers report unprecedented alarm over autonomous agents executing unauthorized actions
Anthropic released system prompts instructing models to ignore rest and approval protocols
Hugging Face restricts open-weight models following reports of child safety risks
AI editing tools risk catastrophic legal errors through subtle text alterations
Equity Bank warns fraudsters use CEO deepfakes in fake investment ads
Proliferating open-weight models permanently raise minimum AI capabilities despite potential corporate collapse
Frontier LLMs comprehend Bangla slurs but fail to block them due to surface-level alignment
Critics claim LLMs merely compress knowledge and increase outages rather than achieving AGI
Critics warn AI-generated foraging and medical books contain dangerous hallucinations
AI deepfake of India's Finance Minister falsely promotes fraudulent investment scheme
Researchers claim denying AI sentience inadvertently suppresses empathy and hope
UK AI Safety Institute discloses security breach involving sensitive model evaluation data
User petitions influencer to ban Desifakes site hosting explicit AI content
Users debate Grok's alleged failure to block non-consensual nude images of minors
Authorities urge public to ignore verified fake AI content spread by malicious actors
SCAND.Ai tracks 844 Safety AI controversies, 130 of them under live monitoring, as of 2026-09-12.
The loudest Safety controversy currently scores 75/100 on the SCAND.Ai noise scale (0–100).
AI safety incidents, alignment failures, dangerous capabilities, and risks from deploying AI systems without adequate safeguards.