There’s a catch-22 here where the wording will put people off if it’s too extreme because they’ll just put all doomsayers in one boat whether the fears are over AI, UFOs or cthulu and then dismiss them equally. (It’s like there’s a tradeoff between level of alarm and credibility).
And on the other hand claims will also be dismissed if the perceived danger is toned down in the wording. The best way to get the message across, in my opinion, is to either have more influential people spread the message (as previously recommended) or organize focus testing on what parts of the message people don’t understand and workshop how to get it across. If I had to take a crack at how to structure a clear, persuasive message my intuition is that the best way to word this message is to explain the current environment, current AI capabilities and specific timeline and then let the reader work out the implications.
Examples
‘Nearly 80% of the labor force works in service jobs and current AI technology can do most of those jobs. In ~5 years AI workers could be more proficient and economical than humans.’
‘It’s impossible to know what a machine is thinking. In running large language model based AI researchers don’t know exactly what they’re looking at until they analyze the metrics. Within 10-30 years an AI could reach a super intelligent level and it wouldn’t be immediately apparent.’
There’s a catch-22 here where the wording will put people off if it’s too extreme because they’ll just put all doomsayers in one boat whether the fears are over AI, UFOs or cthulu and then dismiss them equally. (It’s like there’s a tradeoff between level of alarm and credibility).
And on the other hand claims will also be dismissed if the perceived danger is toned down in the wording. The best way to get the message across, in my opinion, is to either have more influential people spread the message (as previously recommended) or organize focus testing on what parts of the message people don’t understand and workshop how to get it across. If I had to take a crack at how to structure a clear, persuasive message my intuition is that the best way to word this message is to explain the current environment, current AI capabilities and specific timeline and then let the reader work out the implications.
Examples
‘Nearly 80% of the labor force works in service jobs and current AI technology can do most of those jobs. In ~5 years AI workers could be more proficient and economical than humans.’
‘It’s impossible to know what a machine is thinking. In running large language model based AI researchers don’t know exactly what they’re looking at until they analyze the metrics. Within 10-30 years an AI could reach a super intelligent level and it wouldn’t be immediately apparent.’