(Note: This post is aimed at Americans, but similar things may apply in other countries.)
This week is an excellent time to write to your representatives in Congress about halting or slowing AI development, now or in the near future, or at least laying the groundwork to do so.
Because yesterday, over 1,000 AI company employees released an open letter calling for the federal government to do just that. And last week, there was a fairly big news story about an unreleased OpenAI model "going rogue" which probably already caught the attention of your Congresspeople. There's a chance here for your representatives to hear industry experts saying something, the public saying the same thing, and their own gut reaction to recent news saying the same again. Sounds like a persuasive combo if you can make it happen.
Only a few months ago, my sense was that the most pessimistic people about AI safety (e.g., MIRI) wanted to shut things down, while people who worked at OpenAI and Anthropic thought they could just muddle through. But, if you haven't been paying attention, it's not just MIRI people and PauseAI activists calling for a shutdown or at least a slowdown at this point. Anthropic said recently, "If it were possible to effectively slow the development of this technology to give ourselves more time to deal with its immense implications, we think that would likely be a good thing." The AI Futures Project recently put out a policy recommendation they call Plan A that involves a brief halt to global AI development followed by a new governance regime designed to keep things relatively slow and controlled. And now we have this open letter. It seems to me there is an emerging consensus among experts that governments must prepare for, if not initiate, a shutdown or slowdown right now.
Remember that you have one House rep and two senators, so you should be able to email three people. If you don't know who your House rep is, you can look it up here. If you don't know who your senators are, google "senators" along with the name of your state. Your rep/senator will have an email contact form on their website.
Don't worry too much about exact framing or policy details -- whether to tell them to halt AI development now, build the capacity to halt it later, stop the race between companies/countries, prepare to avert a future intelligence explosion, etc. Since there's no concrete piece of legislation in question here, anything in the vicinity will probably get rounded off to the same message. My understanding is that in practice these things will get glanced at by a staffer who will add a tally mark to a spreadsheet to indicate that one more person emailed saying, "AI regulation good."
Feel free to copy any or all of it:
"I'm writing in support of the open letter released on July 28th by AI company employees titled "Pacing the Frontier", which calls for the federal government to prepare to slow the pace of AI development. (Letter available here: pacingthefrontier.com)
The capabilities of AI systems are advancing at an alarming rate while the companies that build them are unable to ensure that the use of such systems is safe, as demonstrated by last week's news that an OpenAI bot hacked a major tech company during what was supposed to be unrelated testing in a controlled environment. Their own leadership and employees are fearful of their coming products, but these companies will continue to recklessly build more dangerous systems as long as they are in a technological race with their competitors in the U.S. and overseas.
As a concerned voter, I ask you to support an international agreement that will end this tech race and allow AI companies to build with safety as top priority. One vision of what such an agreement might look like has been laid out recently by the AI Futures Project under the title "AI 2040: Plan A". (Available here: ai-2040.com)"
