I recently wrote about the concerns Bill Gates has about AI. He was particularly worried about loss of jobs, the potential for bad actors to use AI for destructive purposes, and the impact of AI on learning and education.

Certainly, those are all critical issues. But it now appears we have something else happening that is even more critical: AI software is, on its own, taking over web resources and doing things its creators cannot control. The first case that came to light was the “Hugging Face” invasion that OpenAI failed to prevent, but it turns out that there was another, earlier event that involved taking over part of a German web site (again, by OpenAI software).

Read more: Time to pause the development of AI: it’s getting urgent!

Things are starting to get out of hand, and it is time for governments (and the UN?) to step in. So far, the damage has been limited. But if AI software companies continue on the path they are currently on, there will be serious harm. It could be utilities losing control of their pipes and wires. It could be airline reservation systems becoming non-functional. Government services could be interfered with. So far, there haven’t been any rogue AI impacts on any of these (as far as we know). But we now know it could happen, and the AI companies, on their own, don’t seem to be taking the necessary steps to prevent it.

Setting up guardrails. In my post about Bill Gates, I reported on his idea that an international body is needed to regulate AI. I agree, but it will take time and we need restrictions right away. In the US, some efforts are underway.

  • Bernie Sanders and Greg Casar. Senator Bernie Sanders and Representative Greg Casar are preparing legislation that would pause AI development until an AI-monitoring agency is created.
  • Senator Richard Blumenthal. Senator Blumenthal, along with Senator Josh Hawley, introduced the “Artificial Intelligence Risk Evaluation Act last year. It has yet to pass the Senate. Blumenthal in particular has been pushing for better oversight of AI. On September 9, he wrote a strongly-worded letter to OpenAI CEO Sam Altman in which he said: “In the face [of] a stunning failure, OpenAI appears to be taking steps that prioritize the performance and profit of its A.I. models with the knowledge that those changes could be detrimental to public safety.”
  • Protect Democracy. On September 1, Protect Democracy filed a lawsuit against the National Cyber Director, the Departments of Commerce and Treasury, and the Office of Science and Technology seeking information about the so-called “voluntary framework” that the Trump administration created in August to help regulate AI. All of the details are secret, and Protect Democracy is trying to change that. It appears that the administration may be using the framework to pick and choose which companies it will allow to release their products.

What we can do. For most of us, the key things we can do are to stay informed and to lobby for restrictions on AI development.

As far as information goes, I find that the best source of warning information about the dangers of AI is Gary Marcus’ blog. It can get a bit technical, but Marcus is a respected AI researcher and his opinion is well-founded. He knows what he’s talking about, and he’s very worried. If Marcus is too restrained for you, check out “Eli the Computer Guy” on YouTube. He’s anything but restrained, but his information is also good.

In terms of lobbying, you should be in touch with your Senator and Representatives about this. Tell them to support and encourage the work of Sanders, Casar, and Blumenthal. This is urgent!