I appreciate this disclosure and recognize saying “may provide inaccurate information” is the same as saying “will always provide inaccurate information” if there aren’t ways to prove what is and isn’t accurate. They need to add a disclosure about their anthropomorphization.
SAFETY
-
Open Source Airliner Firmware Requires Certification Compliance
By
–
You certainly can open source commercial airliner firmware.
But you can't deploy it in an airliner without some sort of certification. -
Data Leakage Issues in AI Models: No Clear Evidence Found
By
–
I'm not seeing any clear sign of data leakage issues for any of those models – are you?
-
Gap Between AI Description and Physical Task Execution
By
–
There is a huge gap between regurgitating a description of a physical activity and being able to accomplish it.
-
AI Image Generation Prompt Flagged During Reuse Test
By
–
I passed back one of the prompts used to generate a prior image and it also got flagged.
-
AI Power Redistribution: Tools for Good and Evil
By
–
AI is about to redistribute power like never before.This means that everyone will have access to tools that can be used for good or for evil.
-
Uncommon AI Applications Requiring Physical Intuition
By
–
Almost everything that is uncommon and requires physical intuition.
-
Legal threats prevent transparent answers about AI policy
By
–
“Why will nobody give me a straight answer?” The threat of prison if they do is a major factor. See generally:
-
Regulation Impact on AI Safety: Critical Assessment
By
–
correct. Which is why my original post points out how critical it is to figure out whether advocating for regulation is likely to increase, or decrease, safety.
-
Remote-Operated Autonomous Vehicles: Safety and Cost Effectiveness
By
–
Why does that matter?
The only questions are:
1. Is it safe?
2. Is it cost effective?
If remote operators have to intervene once in a while, so be it, as long as it's safer and cheaper than a human driver in the car. Historically, most AI-assisted services have used a