posted in Technology
'Author: OpenAI': Missouri lawmakers' proposed AI rules were written by AI, metadata shows
www.ksdk.com/article/news/local/author-openai-metadata-contradicts-missouri-lawmakers-denials-about-who-wrote-their-ai-playbook/63-7b808ba8-cce3-4a32-8c55-d22232c61435
Who cares? As long as the law isn’t inherently flawed or unfair to his/her constituents, I could care less. Its not like lawmakers were previously writing law all by themselves. They offloaded that work to legal teams long ago. So if it was written by AI and manually reviewed by a legal team what’s the difference other than a few seconds of compute to generate the text? Training models and building out unnecessary infrastructure is the primary issue with AI.
Because there’s no actual guarantee that it was reviewed by a legal team, as evidenced by the countless examples of politician speeches that left obvious “would you like me to reword as if X” artifacts, and references to completely hallucinated cases in many other court cases recently.
Then the issue should be with laziness and poor quality control by the lawmaker. The output of AI is known to have issues. If an issue is found then the repercussions of the output falls squarely on the lawmaker and should be kept in mind during elections or recalls if applicable.
That is definitely the issue. And this problem is only ramping up in almost all fields.
Yeah. Of course it’s going to be a problem. Just like any tool that allows people to easily offload work. These people are seeking to bring balance back to their life in a world that’s driven their free time to near zero. They need to be taught how to properly apply the tool, cause sorry to tell ya, the genie is out of the bottle. LLMs are here to stay. So we need to learn to use them appropriately rather than trying to outlaw their use in some way.
Until we get a framework for how to use them appropriately as a society we just need to settle on holding people accountable for blindly trusting their output. At the end of the day, you should be responsible for the work you do. And that includes harmful output and/or outcomes of the AI company itself. If the AI is actively spreading misinformation or harmful content the company and its management should be held liable for damages.
Not sure what the context is to warrant this reaction. I am not part of the anti-everything-ai mob.
I fully agree on both points. Llms are here to stay and people need to learn to use them.
That said: “please check the output before you hand it out” seems like such a basic minimum.
i am actually shocked that usually smart people, including school teachers. Don’t practice this minimum of quality control.
But this wasn’t a teacher, this was “the administration” a plural of people of which any could and should have done this minimum.
At that point just hand out empty notebook instead. Its would actually be useable. It would be way less damaging to the image of the school. A School that cant write legible text in officials documents isn’t worth sending my kids too.
Apologies. I read your comment after an inherently negative one and the negative connotation bled into yours. But yeah, we’re in agreance that the issue is ignoring the need for AI literacy of some sort. The bare minimum is being ignored by upper level decision makers trying to step into the roles they manage. In this case administrators thinking they know better than their best teachers and instead writing the curriculum themselves with LLMs.