AI a big collection of fails and a warning

in #writing11 months ago

Playing around with the chat bot chatgpt is probably the best way to learn not to be afraid that it will try to steal your job.

There have already been lots of people showing why this is not an intelligent system, yes it can use language, but it has no smarts, it will for happily create information out of thin air, and furthermore claim that it is valid.
Another thing is the total lack of any intelligence for simple things like spelling while it can spell, asking about how many times some letter appears in a word, it has a hard time with this simple task.

Even though it can spell the word in question.

better example would be

other examples include
this set of failures in a convenient GitHub repo
LLMs failures bing and chatgpt

chatgpt and gpt4 fails to pass medical exams

chatgpt fails economics

there is much more but this should be enough to make my point, but just to add a little counterweight here is a short video about how this AI could be a problem

Danger of training AI to seem human Braxman

Sort:  

Personally I think the danger rather lies within the opinions and thoughts, the so easy possible manipulation on not self thinking people on earth.

this might very well be the most dangerous aspect of it.

Congratulations @iobates! You have completed the following achievement on the Hive blockchain And have been rewarded with New badge(s)

You made more than 900 comments.
Your next target is to reach 1000 comments.

You can view your badges on your board and compare yourself to others in the Ranking
If you no longer want to receive notifications, reply to this comment with the word STOP

Check out our last posts:

Our Hive Power Delegations to the May PUM Winners
Feedback from the June Hive Power Up Day
Hive Power Up Month Challenge - May 2023 Winners List