Identify and assess the risk that AI is misused for cyberattacks, fraud and other crimes or illegal acts and take appropriate measures; using the latest technology and knowledge, work to suppress inappropriate outputs that lead to hallucination, aggravated bias or discrimination and the spread of misinformation or disinformation (including deepfake videos and sexually altered images), and to prevent unintended operation or malfunction; and, given the serious risk of AI-generated misinformation, strive to develop technologies that make AI-generated content identifiable (digital watermarks, provenance management, APIs) and implement them as needed.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.