Hey everyone. Im a noob to AI. I ran thru this how-to a couple of times, and have an api, which was created in AWs api gateway. https://builder.aws.com/content/2nF0CBYXKZdaSaCG758grnAkSSH/deploying-ai-agent-apis-on-aws-the-easiest-way The api looks like this: https://zz9v42px6a.execute-api.us-east-2.amazonaws.com/prod/chat I tried connecting to it like this: wscat -c wss://zz9v42px6a.execute-api.us-east-2.amazonaws.com/prod/chat error: Unexpected server response: 403 <<<<my error Ive since deleted the stack Any thoughts? I have no idea what Im doing again : ) What I would like to do is finish this stack with creating a front-end to communicate with the agent. The agent uses a generative AI Large Language Model (LLM) on AWs Bedrock.

AI can automate almost anything—but if we don’t secure it, we create more problems than we solve. So here’s a challenge: Pick one small AI process you’ve implemented (or plan to) and map out what could go wrong. ✅ Could someone spoof the inputs? ✅ Is data encrypted at rest and in transit? ✅ Does it log activity for review? Then… share your answers below. I’ll be here giving feedback and suggestions 💬 Let’s get used to thinking like builders AND defenders.

Meta has refused to sign the European Union’s code of practice, although the AI Act (effective August 2, 2025) is legally binding, whereas this Code is voluntary. It will have implications to how we approach AI security with our clients. Meta is currently the only major U.S. tech firm declining to sign the code, OpenAI and Mistral have committed to signing; Microsoft is “likely to” (Reuters). Thoughts?