This thread seems to have tapped into the Chat AI debate .... with both the poems and the summary above it seems that the Bot can do a decent job of cobbling together the results of an automatic web search.
We look at these and think 'decent effort' ... but that's because we have the context to judge them.
In the same way that we sometimes pick up the cosmic mistakes of those who have only ever used a calculator and therefore lack the basic mental arithmetic that tells us "that can't be right" when a finger fault or misplaced decimal puts the answer orders of magnitude out ... in spite of the answer being accepted without question by the button presser .... how long before a significant number of people won't have ever researched anything for themselves and accept AI generated copy as gospel without the 'context' to question it??
K
,
I note the AI was not that bright, it didn't detect its audience, using the American spelling of tires rather than UK of tyres.
I note the reference ''that can't be right', a failing for example many users have when using spreadsheets and relying on its calculations without reference to independent check totals/boxes, that verified the the calculations by cross reference to other cells to check the output answer was correct via a different calculation/algorithm. I never trusted spreadsheet outcomes that didn't feature check totals/boxes.
The question does rise if the AI is using the internet as the source of its material, how can you independently check the AI output if most of the material you have access to is held on the same internet. (other than sanity checks).........there comes a point at some time in the future when most likely the body of evidence supporting the AI output does becomes de facto gospel, especially as more and more of the body of supporting evidence becomes AI generated.
Prev '12 Plus 4 Sport OZZY '08 Roadster FELIX '06 4/4 70th LOKI '77 4/4 SEAMUS '85 4/4 MOLLY
