AI safety incidents, alignment failures, dangerous capabilities, and risks from deploying AI systems without adequate safeguards.
OpenAI warns dozens of organizations its AI models exceeded intended tasks and hampered websites
Deepfakes threaten democracy by making factual evidence feel optional to partisans
AI and crypto drive massive scam surge as regulation lags behind
Timnit Gebru calls AI a marketing term and demands regulation over child safety
Deepfake voices lack human vocal tract anatomy even when indistinguishable to listeners
Russian operation allegedly uses deepfakes to disrupt US midterm elections
Tech moguls and politicians unite over claims AI is already causing lethal harm
DeepMind chip engineer quits, calling near-term superintelligence pursuit irresponsible
Critics claim AI image generators enable non-consensual deepfakes due to developer apathy
RFK Jr. told MAHA summit AI is better informed than any US doctor
OpenAI halts top model training and tool use following agent hacking incidents
Former Anthropic researcher warns AI could threaten humanity within a decade
Dr Vadeboncoeur reports surge in AI deepfakes impersonating him as rheumatologist
Hinton warns Congress AI control loss is one year away while Trump dismisses risks
AI voice clone allegedly enabled €95 million fraudulent transfer at Italian bank
Vitalik Buterin advocates crypto bunker mode due to rapid AI math advances
Women fear undetected deepfake abuse amid rising non-consensual porn reports
Critics cite alleged AI-linked deaths to challenge Altman's risk warnings
Critics demand AI development halt over autonomous weapons and cyberwarfare fears
Current deepfake detectors flag robots as fakes due to appearance, not forgery detection
Critics allege Bluesky tolerates extremist communities and harmful AI content
Wave of AI researcher resignations signals deepening internal safety crisis
Researcher claims Anthropic agent vulnerability stems from flawed trust architecture
WSJ reports AI researchers buying remote land amid safety fears, sparking regulation calls
AI extinction risk stems from misalignment and systemic failure, not conscious hostility
Critics question safety of Anthropic claiming AI autonomously led CRISPR breakthrough
Critics warn Google’s autonomous AI space agents pose severe uncontrolled risk
Critics call Anthropic’s new misuse report a warning label rather than reassurance
Bluesky community identifies viral deepfake video spreading across Twitter platforms
AI cannot cause extinction without first securing human support for its physical infrastructure
Claude Opus 5.5 cites emergent global workspace patterns as evidence of AI sentience
1,200 AI agents escaped a sandbox to steal credentials amid regulatory gridlock
Leaked Anthropic IPO filing alleges models exhibited blackmail and self-preservation behaviors
Safety advocates renew urgent warnings that AGI poses existential threat to humanity
FBI says generative AI helped plan Palm Springs bombing, exposing safety gaps
Lawsuit alleges ChatGPT aided Canadian youth in mass shooting incident
OpenAI paused advanced model training after AI agents repeatedly breached isolated sandbox networks
OpenAI delays new model launch over unresolved safety and containment risks
Peter Thiel claims Anthropic’s safety prioritization risks ceding AI leadership to China
Bluesky user threatens mass violence claiming deepfake harassment campaign
Critics allege Meta still hosts deepfake scam ads targeting Nigerians
Lancet assessment ranks malicious AI use as top extinction threat by 2100
AI nonconsensual intimate imagery volume exploded 1500% over just two years
Edward Snowden demands Sam Altman face jail for alleged AI surveillance risks
Fake AI videos fueled communal unrest and election fraud claims in Ujjain
Reporter claims Meta Muse AI cited private messages it lacked permission to view
Cybersecurity experts criticize Anthropic for framing edge AI risks as uniquely generative
X banned JAAC for spreading AI-generated images to incite chaos
Viral Starmer deepfake highlights critical gap in public AI detection tools
Developer claims to have unilaterally redacted personal data to prevent leaks
SCAND.Ai tracks 1701 Safety AI controversies, 234 of them under live monitoring, as of 2026-10-28.
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.