generative AI models: Thanks for the info, hush with the source



While New York Times sues OpenAI and Microsoft for feeding their generative AI beasts fodder from the organic food trough of the Grey Lady, we ponder over what happens to the news that is not fit to print, but comes out of ChatGPT and Bing Chat smelling of NYT anyway. Apart from the newspaper making a lot of money if it wins the case, there is a chance that those emulating the two most famous generative AI models will find a way to tweak language learning models by leaving no – pardon the pun – paper trail.

As every college plagiarist worth his or her salt knows (by picking this knowledge up from somewhere else), nearly nothing is original or new, and certainly no information is anyone’s monopoly. Something that many Hindi film scriptwriters are familiar with. So, the question becomes of acknowledgement. Which is another way of stating the amount one fixes to a bill of information-sharing. While purists may argue that there is a price to be paid for quality legwork done to obtain information, the real world increasingly shows the world caring about who ‘told them so’ or what they are told. All that seemingly matters is they are told something. Which means NYT may believe only in news that is fit to print, but chatbots need not be so restricted. Since the public, dear reader, doesn’t want restrictions of quality or veracity either.



Source link