AIThe debate over trust and monitoring in AI models
One user warns that systems are becoming “unmonitorable.” A critic argues that the user’s company chooses not to monitor its models and to develop ones that are harder to monitor.
TLDR
One user clarifies that their complaint is about systems quickly becoming “unmonitorable” while people take them at their word. Quoting that comment, another user alleges that the original poster’s company chooses not to monitor its models and chooses to develop harder-to-monitor ones. The critic argues that the company could stop making those choices.
Combined views
2K
1 Source, first seen 6h ago
The debate over trust and monitoring in AI models
One user warns that systems are becoming “unmonitorable.” A critic argues that the user’s company chooses not to monitor its models and to develop ones that are harder to monitor.
TLDR
One user clarifies that their complaint is about systems quickly becoming “unmonitorable” while people take them at their word. Quoting that comment, another user alleges that the original poster’s company chooses not to monitor its models and chooses to develop harder-to-monitor ones. The critic argues that the company could stop making those choices.