Daily review 14/09/2026

3 comments on “Daily review 14/09/2026”

Comment numbers have changed to permanent references. Older quoted numbers may no longer identify the same comment.

1. 1

   I think weka might enjoy this looong longform article, a chat with computational linguist EM Bender. It's an expansion in a new direction on her robot post. Bender identifies another danger of Large Language Models, a societal harm, apart from that of ai malicious hackerware.

   'We go around assuming ours is a world in which speakers — people, creators of products, the products themselves — mean to say what they say and expect to live with the implications of their words…. we’ve altered the world. We’ve learned to make “machines that can mindlessly generate text…But we haven’t learned how to stop imagining the mind behind it.” '

   “…applications that aim to believably mimic humans bring risk of extreme harms….Work on synthetic human behavior is a bright line in ethical Al development, where downstream effects need to be understood and modeled in order to block foreseeable harm to society and different social groups.”

   Bender was one of three authors of a 2020 Stochastic Parrot paper, which pointed out the obvious: LLM models that have been trained on the English, white, western male canon, not to mention a whole lot of internet sleeze, will display an very obvious bias. The other two women authors lost their jobs at Google when they refused to retract the paper when ordered to by their bosses.

   There's quite a lot more in this article on risks and uses of ai in a social context, and Bender's philosophical position in comparison with techbros, who are definitely not humanists.

   ‘…Dennett, the philosopher of mind, [is] even more blunt. We can’t live in a world with what he calls “counterfeit people.” “Counterfeit money has been seen as vandalism against society ever since money has existed,” he said…”Counterfeit people is at least as serious.” ‘

   https://nymag.com/intelligencer/article/ai-artificial-intelligence-chatbots-emily-m-bender.html
   1. 1.1

      That was a March 2023 article, but still veey pertinent.
      1. 1.1.1

         Even more pertinent in view of the 'progress' in AI and the deepening concerns from within the industry

         …. LLM models that have been trained on the English, white, western male canon, not to mention a whole lot of internet sleeze, will display an very obvious bias

         Massive bias.

Comments are closed.