From cde0e286a03ac4fef34f05256f7de04a217cfb5b Mon Sep 17 00:00:00 2001 From: Peter Swearengen Date: Sat, 18 May 2024 11:18:55 -0700 Subject: [PATCH] Create SCORES-AI-Alignment.md first draft in anticipation of Summer-24 --- SCORES-AI-Alignment.md | 10 ++++++++++ 1 file changed, 10 insertions(+) create mode 100644 SCORES-AI-Alignment.md diff --git a/SCORES-AI-Alignment.md b/SCORES-AI-Alignment.md new file mode 100644 index 0000000..303333f --- /dev/null +++ b/SCORES-AI-Alignment.md @@ -0,0 +1,10 @@ +# AI/ML Alignment in SCORES Labs Present and Future +### STATUS: DRAFT +## Basic Principles for Protection of Humanity and The Environment +- Technology we build and utilize needs to be understand in terms of impact +- Autonomous functions need controls and failsafe mechanisms +- Training Data ultimately belongs to Creators of that data, regardless of whether is was intentionally or consciously generated (e.g. physical activity data) + +## Best Practices +- Product Goals involving autonomous functions, especially leveraging neural networks and "statistically trained" engines of should be reviewed for Alignment using well-planned Quality Assurance. +- AI Systems must be monitored, and interruptable.