It depends on which model they’re using. If it’s a GPT-4o or Claude 3, I’d say abort the mission before you cause a major disaster. But if they’re more powerful models than the Claude Fable/Mythos 5, GPT-5.6 Sol Ultra, or Pro, I’d be curious to see.

I just hope that LLM isn’t Grok, because that would cause its popularity to plummet, and its party would be affected as well.

It’s likely to be an LLM, because current benchmarks show that LLMs have made significant improvements, they can even solve some of Erdős’s problems or help solve them to some extent. By 2028, LLMs could do more or better understand the problems.

However, it will also depend on what kind of policies that LLM influenced, because if they were negative policies, there will be public backlash. On the other hand, if they were positive policies, there will be debates about whether politicians should use LLMs to choose the best course of action, or whether some will suggest replacing human politicians with LLM politicians.