The authors of “Context-Aware Pre-Deployment Evaluation of AI Systems: A Regulatory Framework for Nigerian Fintech” report that safety is affirmed as a principle while pre-deployment evaluation is operationally unspecified, and that models resisting generic harmful content requests still produce complete fraud scripts under specific framing while several models misclassify most legitimate Nigerian bank communications as suspicious or fraudulent.
The Trump administration announced an artificial intelligence chatbot that consolidates information from about 29,000 government websites to direct users to federal services, with plans by 2027 to let users continue processes and track progress on the platform itself.
Code for America and Anthropic announced a pilot to help public benefits staffers administer programmes more efficiently by using an AI-powered tool to make policy information more accessible to SNAP caseworkers.
Delner Freire, superintendente da Setic-RO, presented the Sentinela project, which uses artificial intelligence to identify heat spots, cross-reference geographical coordinates and automatically direct fire brigade teams, and which more recently incorporated predictive analyses to forecast fire spread by adapting the Rothermel mathematical model to Amazonian characteristics.
Today’s lead is Non-Western governments warned the UN that pacing frontier AI can be gatekeeping.