Nuked 
Hello followers! Let’s dive into the fascinating world of Microsoft’s AI code of conduct—here’s everything you need to know to be both responsible and clever with AI tech.
Microsoft’s AI Services have a new set of rules that ensure applications built with their tools are used ethically. This document replaces older guidelines for generative AI, vision APIs, and speech services, aligning better with current AI laws, like the EU AI Act.
Customers deploying these AI tools must implement strong measures to prevent fraud, reveal when content is AI-generated, and keep human oversight active. Transparency about AI-generated voices, images, or videos is key to avoid deception.
There are strict restrictions on how AI can be used, especially to avoid harm—like not creating deepfakes, non-consensual content, or engaging in activities that violate laws, promote discrimination, or cause psychological or physical damage.
Using AI for harmful purposes—such as child exploitation, non-consensual content, or promoting violence—is completely prohibited. Similar rules apply to content that promotes terrorism, discrimination, disinformation, or malware.
For AI systems with autonomous decision-making, there must be controls and transparency—users should understand how decisions are made, and systems should include measures to detect failures or anomalies.
Specific rules govern vision and face recognition APIs, requiring safeguards against misuse like biometric attack or bypassing security features. Voice AI services, such as text-to-speech, need user consent, clear disclosures, and options to opt out, especially for external users generating voice content.
Limited access services are available to approved customers under strict terms, ensuring responsible usage. If misuse is suspected, users are encouraged to report abuse using designated tools.
Overall, this Code emphasizes diligent, transparent, and lawful use of AI, with extra rules for technologies like face recognition, voice synthesis, and autonomous systems to prevent harm and ensure trustworthiness.
