Ask, then verify
Each answer starts with the answer, then links to the tools and guides that go deeper, and to the sources we used.
Microsoft's AI Red Teaming Agent docs show a model refusing a direct harmful request and answering it once the characters are flipped. AI red teaming hunts for that kind of failure before release; it finds risks, and systematic measurement then tells you how common they are.
Customer records and credentials reach language models through prompts, uploads and retrieved documents, and OWASP warns that data users provide can later appear in a model's output. Controls detect and redact or tokenize that data, limit access, and show which AI tools staff use.
Guardrails decide what reaches your model and what reaches your users, yet AWS describes even Bedrock's sensitive information detection as probabilistic. Choose one by where it runs and how it performs on your own prompts, and pair it with least privilege and output handling that treats model text as untrusted.
Every MCP server you connect can feed instructions to your model and act with the permissions you grant it, and a server you approved can change its tools later. Treat each Model Context Protocol (MCP) server as code that runs against your data: scan it, pin it, limit its scopes and keep a human in the loop.
Loading a PyTorch pickle file from an untrusted source can run an attacker's code on your machine, so treat downloaded models like downloaded code. Safer formats, scanning, signing and an inventory of what you run reduce the risk without removing it.
A web page or file your LLM reads can carry hidden instructions that make it leak a private conversation or misuse its tools. OWASP ranks prompt injection first (LLM01) among LLM application risks and says it is unclear whether fool-proof prevention exists.
Superintelligence does not exist yet, and no tool page in this directory documents a way to secure it. Google DeepMind's published approach to AGI safety and security still relies on controls that security teams run today, such as access control, monitoring and red teaming.
Terms are defined in the glossary, and the risks behind these questions are explained in the OWASP Top 10 for LLM Applications.