This is the result of a half-day research sprint on the recently-introduced amendment to institute a 10-year moratorium on state-level AI regulations in the current budget reconciliation bill, with a focus on figuring out "is it likely to survive the Byrd Rule".

It seems quite obviously in violation of the Byrd Rule because it violates at least 2 of the 6 tests^[1]:

• It has no budget effect, i.e. does not change outlays or revenues. (Probably this is not actually true; I think on net we would expect it to somewhat increase revenues via 2nd/3rd order effects, but in the relevant legal sense, it's not an explicit additional tax and it's not an additional explicit expenditure.  Also, the CBO estimate for that entire section is -500m, which is the full sum of the $500m it appropriates to modernize and secure Federal information technology systems through the deployment of commercial artificial intelligence, the deployment of automation technologies, and the replacement of antiquated business systems in accordance with subsection (b))
• Even if it did have a budget effect, it would be incidental to the non-budgetary component of the provision, which is clearly meant to be a "policy change".

There is a "Byrd bath" that reconciliation bills go through, directed by the Senate Parliamentarian, before the bill is taken up by the Senate.  The Parliamentarian & Senate staff identify the sections that seem like obvious violations of the Byrd Rule and try to work with whoever drafted the relevant section of text to redraft or amend it so that it doesn't violate the Byrd Rule (if possible; sometimes they're just deleted wholesale).  It's not clear to me what percentage of violations get scrubbed out at this step, but various LLMs think it's most of them.

If it does still somehow make it through that step, Senators can raise a "point of order" against subtitles, sections, lines, or even single words in the bill.  Other Senators can motion to "waive" a point of order; these waiver motions require a 3/5 majority vote (60 Senators) to pass.  Separately, the Senate chair (currently JD Vance) can rule against the point of order without another Senator initiating a motion to "waive" it.  Historically, the chair has deferred to the Parliamentarian's advice about whether the text being objected to is extraneous or not; as far as I can tell there haven't been any instances of the chair ignoring the Parliamentarian (though it's not clear how likely we'd be to know if it happened).  The Parliamentarian is purportedly non-partisan.  (The current Parliamentarian is a democrat.)  The chair's rulings can be appealed; those appeals also need a 3/5 majority to overrule the chair.  (I haven't figured exactly out how often the chair has ruled against points of order absent motions to waive, but I'm pretty sure it's quite rare, see below.)

There have been ~85 points of order raised since 1985, across 23 reconciliation bills.  73 were sustained, and 12 of them "fell".  Of those, 11 were before 2000; after that, only a single section from a bill in 2005 that had a point of order raised against it has survived.  There were 69 waiver motions and 60 of those were rejected.  From that we can infer that 9 waiver motions weren't rejected, so maybe 3 points of order had the Senate chair rule against them, without a successful appeal overturning that ruling.

I currently think it's pretty likely that the Byrd bath will remove the offending provision before the Senate takes up the bill.  If it doesn't, then I think it's still quite likely that a Senator raising a point of order against it would succeed in getting it struck from the bill, though having the provision survive the Byrd bath would be a bit of evidence that the Parliamentarian didn't consider it a violation (though not fully dispositive).

The Byrd bath stage doesn't seem amenable to lobbying or public pressure; they're basically already trying to do the thing you might want them to do there.  Thus the most effective pathway I can see to affecting the outcome is to preemptively encourage Senators to raise a point of order against^[2] the moratorium text, the usual way (personalized phone calls to your Senator's legislative staff seem to be the best, followed by personalized emails, following by form templates).  There are already multiple groups paying attention to this.

Primary research sources:
 

1. https://d1dth6e84htgma.cloudfront.net/Subtitle_C_Communications_4e3fbcc3bc.pdf (the section of the bill itself)
2. https://chatgpt.com/share/68365b19-9340-800e-944c-19b3cd73e11a
3. https://sgp.fas.org/crs/misc/RL30862.pdf (most of the concrete historical info re: previous points of order and waiver motions comes from here)
4. https://www.cbo.gov/publication/61420 (CBO estimates; you want Title IV, Sec 43201)
5. https://chatgpt.com/share/68365c2d-ba28-800e-a098-cc20dc019a49

Many other LLM queries omitted.

1. ^{^}

   • it does not produce a change in outlays or revenues or a change in the terms and conditions under which outlays are made or revenues are collected;
   • it produces an outlay increase or revenue decrease when the instructed committee is not in compliance with its instructions;
   • it is outside of the jurisdiction of the committee that submitted the title or provision for inclusion in the reconciliation measure;
   • it produces a change in outlays or revenues which is merely incidental to the nonbudgetary components of the provision;
   • it would increase the deficit for a fiscal year beyond the “budget window” covered by the reconciliation measure; and
   • it recommends changes in Social Security.

   Source

2. ^{^}

   If you want the opposite, you might just ask them to... not raise a point of order?  I think trying to get a motion to waive to 60 votes seems pretty hopeless.