33 Comments

Notably, the study found a statistically significant rate at which legislators responded less to AI generated content than actual human generated content.

Ok, so let's do the same study with your rules. Cornell students generate the letters and solicit actual constituents in the legislator's district to hand wright letters and send emails.

Now what?

So legislators were the subject of a study. On average they each got 4.5 letters.

Is your problem with AI generated content?

I would think foreign actors and activist have been doing this for years without AI.

Expand full comment

My point is that if I use an AI-generated letter to send to my legislator, or anyone, I should take responsibility for its contents just as if I composed it, by not trying to be anonymous. If that habit were followed perhaps more folks would read what they post, especially if they didn't actually write it.

Expand full comment

It’s a felony to pretend you are an officer of the court. A US Congressman cannot be arrested going to or coming from a session of Congress (interesting but true still, I believe). To make false claims to the FBI is also a felony I believe; it is a crime to do so in a police investigation. Isn’t there a parallel application available here for professors and students who want to “mess with conservatives” in such a willful and contemptuous manner?

Expand full comment

Step one would be for the GOP (or preferably someone more competent like Americans for Prosperity) create a tool that validates constituents and compares them to things like voting record as well to quickly ID phonies. There are plenty of available services that use public records to validate people and address without invading privacy.

I'm sure the counter would be for the perpetrator of the fraudulent letters to send letters on behalf of actual voters who know nothing about it. So step two should be to have all these letters scanned in and have our own AI connect them together to point to the source and charge those people with fraud.

Expand full comment

You'll hardly ever hear me saying that we need a law, but AI deception is already out of control and bordering on dangerous. I like DeSantis, but his Super PAC's ads (linked below) with Trump's AI voice reading his own "Truth" and with the photos of Trump hugging Fauci are simply wrong. An image or video fabricated by AI needs a watermark. Same with audio and written material. If we can't believe our own eyes or ears, how can we ever learn the truth?

https://www.youtube.com/watch?v=LKQiTpiPN7I

https://twitter.com/DeSantisWarRoom/status/1665799058303188992?s=20&utm_source=ActiveCampaign&utm_medium=email&utm_content=AI+Infiltrates+the+2024+Race&utm_campaign=AI+Infiltrates+the+2024+Race

Expand full comment

It's being driven by the Father of Lies. That's a biblical reference, not a Fauci reference ;)

Expand full comment

This sounds like a really dumb 'study'. The results are obvious, yes, handwritten notes will get noticed and will result in action more frequently by a representative than typed notes or emails. This obvious finding can only be used for nefarious purposes. It should be illegal.

Expand full comment

sigh. This is why we can't have nice things.

Expand full comment

"It's ALWAYS something!" - Rosanne Rosanna Danna

Expand full comment

If we all got back into the habit of taking responsibility for the opinions we share, by using our real names, would it matter if the original was written with AI assistance? Penny N. Gaffney

Expand full comment

Not the point, you as a real constituent using AI to write your letter is fine, same as Erick sending out one for us to use but faking a constituent should be illegal. It is also a waste of taxpayer money if they send out responses in wasted time, envelopes etc. in replying to say nothing about swinging a vote one way or another.

Expand full comment

Is the professor AI?? ...Since he's not responding 🤪🤣

And yeah, there should be a law against that for sure.... AI disturbs me. A lot.

Expand full comment

Erick maybe the AG of West Virginia could apply a little pressure for the professor to respond

Expand full comment

We don't need a new law. We already have laws under the books for commiting mail fraud. That is what it is. The monetary is the time the staffers had to deal with the letters.

Expand full comment

but is the email ones "mail fraud", and i doubt could get DOJ to prosecute it as mail fraud either way.

Expand full comment

This kind of deception has been used by academia for years. In the eighties I worked at a university and was walking by an office that appeared to be dark. As I got even with the door I saw a “black light” on the desk and a phd student I knew working under it. He was taking surveys returned from farmers in the state and checking for serial numbers which were written in “invisible ink “ on the reverse side. I questioned him and he said that he never said that the surveys were anonymous - so he marked them all before they went out and now was “converting” the serial numbers to human readable...

Expand full comment

Wow! Is it any wonder so many of us are so cynical?

Expand full comment

Sounds like a significant mis/dis-information paradigm shift to me. Our new Pravda has arrived... 🤮

Expand full comment

Democrats have a bad policies and they aren’t popular. If anyone actually read what the dems want. Killing kids. Higher taxes and eliminating voting rights. They would be appalled.

But what the dems are good at is politics and are very sneaky at it too.

In Feb 2020 all was bleak for the dems. Economy booming trump up in all the polls. Then the pandemic hit and they took full advantage of it with mail in ballots. And they won. They did the same thing in 2022.

Republicans have good policies and know how to run a free country. Dems know how to scheme and win elections while implementing socialism.

They hate elections.

This AI scheme is just another way for them to cheat elections to get what they want. They have to. Because as mentioned if people really knew the truth they would never vote for a democrat ever again.

Expand full comment

ChatGPT/AI is Skynet and it needs to be destroyed before it destroys human civilization.

Expand full comment

And there goes the last shred of representative government.

Expand full comment