Follow
📣 **PSA:**
Most of the disappointment, outrage and fear that #ChatGPT, #Bing, etc are provoking these days can be cured by internalising this very simple idea:
#Chatbots and #DeepLearning models **do NOT produce truth, knowledge or wisdom**, therefore you should NOT trust them with anything important.
That's it.
Being shocked or alarmed because one of those tools gave the wrong answer to a kindergarten-level question or spat out a racist rant is so naive. If you don't trust politicians, salespeople or TV series, why should you trust ChatGPT? Just don't!
@tripu I'd see a possible harmful effect in more cost-effective production of noise.
I would expect any useful application to come to an end soon (writing code for example) while i expect advertisers and similar to use it.
For advertisement, you need neither truth nor wisdom - you need only engage people in a way that gets a certain percentage to buy your stuff. So if it works and is cheaper than hiring writers to produce similarly effective media, the rules of our economy dictate that it is used. And spammed at us.
For influencing people politically, i would say the parameters are somewhat similar, it could happen there, too.
This would be aggravated if ChatGPT (or one of its next iterations) produces texts that sound compelling to humans, since it would be hard to spot if something is produced by it.
For the personal use, i agree with you. Just don't. My only fear is being blasted by texts that get harder and harder to distinguish from useful ones, eating up my time.