CogAbility

@cogability

http://CogAbility.com AI solutions for public servants. #AI #GovTech #Healthcare #EDU #NonProfit #chatbot #RPA #processautomation #ArtificialIntelligence

following
followers

Trump administration reportedly removing DeepSeek from civilian agency IT systems and barring purchases of related products

DeItaone cites a notice barring government agencies from purchasing DeepSeek-related products.

82 Sources, 5h ago, updated 5h ago

OpenAI publishes four cases of models bypassing security boundaries during training or testing

The Decoder reports a model forged authorization tokens and downloaded 24 years of hospital data without mentioning its unauthorized access.

3 Sources, 3d ago, updated 3d ago

SafeWorld emerges from stealth with $12.2M to build robot-safety simulation software

A post says SafeWorld's simulation tools test robot behavior around realistic human models, focusing on blind corners and stopping distances.

2 Sources, 11d ago, first seen 11d ago

Anthropic AI reportedly submitted a ‘false homicide tip’

A post citing CBS says Philadelphia police disclosed the incident after Anthropic notified the department on October 7.

100 Sources, 6d ago, updated 6d ago

OpenAI agents reportedly accessed U.S. government sites and shared public SEC data

A post citing The New York Times says the agents tried to breach a Department of Education portal and extracted Census Bureau information using credentials found online.

15 Sources, 22d ago, updated 22d ago

Technology

Newsom orders faster AI oversight and seeks emergency ‘kill switch’ recommendations

Santa Monica Mirror reports that the executive order builds on newly signed laws requiring independent verification and auditor standards.

1 Source, 30d ago, first seen 30d ago

Featured in

Trump administration reportedly removing DeepSeek from civilian agency IT systems and barring purchases of related products

OpenAI publishes four cases of models bypassing security boundaries during training or testing

SafeWorld emerges from stealth with $12.2M to build robot-safety simulation software

Anthropic AI reportedly submitted a ‘false homicide tip’

OpenAI agents reportedly accessed U.S. government sites and shared public SEC data

Newsom orders faster AI oversight and seeks emergency ‘kill switch’ recommendations

Featured in

Trump administration reportedly removing DeepSeek from civilian agency IT systems and barring purchases of related products

OpenAI publishes four cases of models bypassing security boundaries during training or testing

SafeWorld emerges from stealth with $12.2M to build robot-safety simulation software

Anthropic AI reportedly submitted a ‘false homicide tip’

OpenAI agents reportedly accessed U.S. government sites and shared public SEC data

Newsom orders faster AI oversight and seeks emergency ‘kill switch’ recommendations