---------- Forwarded message --------- From: Keith Burrows <keithbphysics@gmail.com> Date: Wed, 3 Sept 2025 at 17:35 Subject: Re fixing AI etc Another try To: Paul Cooper <paul@project-143.com>
Hi Paul, sorry if this is doubled up, the first try included your smsmt email and was rejected.
Thanks for your various emails about AI. The “Five things to fix” seems to make sense to me, despite my lack of use of much AI.
But I do have a worry about AI that I haven’t seen addressed much - or at all. If I ask AI about climate science, for example, where does it find the answers to my question? There is an awful lot of “rubbish science” out there. Can it tell the difference? I doubt it goes straight to Nature or Science but rather goes to “second hand” sources such as New Scientist or Scientific American (hopefully). But what about all the pseudo science out there? Can it really distinguish that from the real science?
Then that leads to my next problem. If everyone just goes straight to AI, will those “second hand” sources dry up from lack of use? I know that The Conversation (a source I very much respect) has expressed concern about lack of support as a result of easy AI access.
Clearly there are uses of AI that make sense, but I do have to worry about the long term effects…?
On a completely different sort of AI (Actual Intelligence!) I came across a very old Nexus issue from Carey (1973 I think) which included the article we wrote . I’ve attached it. At least they did eventually let in the other half of the population and did get rid of silly caps. But I’m not sure much else has changed .
Cheers, and keep up the good work! Keith
