The Responsible AI Safety and Education Act (RAISE Act) is a New York AI safety bill that requires AI developers to produce safety protocols to address the most severe risks (and then follow those protocols), and creates whistleblower protections for AI company employees.
This market will resolve based on the furthest stage the bill reaches by January 1, 2026. Possible outcomes include:
Not approved by the legislature
Approved by the legislature but did not become law (e.g. because the governor vetoed it)
Has become law
Substantial amendments to the bill (e.g. ones that significantly water it down) will not change the resolution of this market, unless the amendments are so substantial that it's a fundamentally different bill (as happened with TRAIGA in Texas). Sorry that this isn't a clear-cut resolution criterion -- I'll make judgments on a case-by-case basis, though I suspect that it won't end up being relevant.
Federal preemption is not relevant to the resolution of this market. For instance, if the governor signs the bill but federal preemption prevents it from taking effect, the market nevertheless resolves "Has become law".
Update 2025-12-06 (PST) (AI summary of creator comment): Regarding amendments to the RAISE Act:
Substantially weakened versions of the bill will still resolve to the appropriate outcome (e.g., "Has become law" if signed)
A NO resolution would only occur if the bill becomes fundamentally different (a complete change to what the bill does, not just a weakened version)
There remains inherent ambiguity around the exact boundary for extreme weakening cases
Update 2025-12-11 (PST) (AI summary of creator comment): If the RAISE Act is replaced by SB 53 near-verbatim, the market will resolve to "Not approved by the legislature" (NO).
Update 2025-12-19 (PST) (AI summary of creator comment): Resolution may be delayed until January when the bill text becomes public, as the creator needs to assess whether the signed bill has sufficient overlap with the original RAISE Act to count as the same bill (versus being fundamentally different). The creator is seeking trader input on how this should resolve.
Update 2025-12-20 (PST) (AI summary of creator comment): The final bill text will not be available until January. The active amendment (B) visible on the NY Senate website is not the final text that was signed.
1,000People are also trading
I'd like to see some traders (ideally ones who don't have large positions, but everyone is welcome to make arguments) weigh in on how this should resolve. In the meantime I've sold my position so that I don't have an incentive to resolve it one way or another, and I won't trade on this market anymore.
I think it's not clear, because my understanding is that the bill that was signed didn't have all that much overlap with the original RAISE Act.
It's possible that resolution would only come in January, once the bill text is public.
@EricNeyman Is the active amendment (B) not the final text?
https://www.nysenate.gov/legislation/bills/2025/A6453/amendment/B
@EricNeyman https://www.governor.ny.gov/news/governor-hochul-signs-nation-leading-legislation-require-ai-frameworks-ai-frontier-models
Flashback: Prior to Hochul signing the bill, tech and AI lobbyists were able to negotiate changes so it more closely resembled California's SB 53, arguing that state bills should be uniform for industry certainty.
But the bill sponsors secured some changes to strengthen requirements around reporting critical safety incidents in the final version.
Bores said in the release that negotiations this week among himself, Hochul and co-sponsor state Sen. Andrew Gounardes were successful.
"We defeated last-ditch attempts from AI oligarchs to wipe out this bill and, by doing so, raised the floor for what AI safety legislation can look like.
"And we defeated Trump's – and his donors – attempt to stop RAISE through executive action greenlighting a Wild West for AI."
https://www.axios.com/2025/12/19/new-york-ai-safety-bill-hochul
So not a near-verbatim replacement
I just want to clarify this part of the resolution criteria:
Substantial amendments to the bill (e.g. ones that significantly water it down) will not change the resolution of this market, unless the amendments are so substantial that it's a fundamentally different bill (as happened with TRAIGA in Texas).
If the RAISE Act is really substantially weakened, this still resolves YES. A NO resolution would entail a complete change to what the bill does, as if it was a totally different bill, rather than a weakened version of the same bill. There's still an ambiguity around cases where the bill is extremely weakened (if the final version of the bill is literally empty, this will resolve NO), but there will inevitably be ambiguity around where the boundary is exactly, sorry.