News and Info for NYS Council Members, 8/20

August 20, 2025

Many NYS Council members have been wondering if and when our NYS will promulgate regulations on the use of Artificial Intelligence (AI) in healthcare settings. It is likely there will be regulations coming from state agencies shortly.  In the meantime, we wanted to share some information about the RAISE Act, passed by both houses of the NYS Legislature at the end of session in June, 2025.  The legislation is now awaiting final action by Governor Hochul.
 
So, before we get specific on the use of AI in MH and SUD settings,  here’s some useful (and plain language) information from The Transparency Coalition regarding the legislation that awaits final action by the Governor.  Note: The legislation is primarily focused on large AI companies and their development of AI models.  The intent is (according to the legislative write up) to avoid or mitigate development of AI models that could result in critical harm.  
 
RESOURCES:
 
-Here’s a look at the legislation awaiting final action:  https://nyassembly.gov/leg/?bn=A6453A&utm_source=substack&utm_medium=email
 
-And here is a useful analysis of the bill (in plain language) from The Transparency Coalition:
https://www.transparencycoalition.ai/news/guide-to-the-raise-act-the-new-york-responsible-ai-safety-and-education-bill
 
-And finally, here is an article from Dan Clark in today’s Capitol Confidential on this topic.  In the days to come we will get more granular with a look at what might be coming from the state agencies in the days to come.
 
 

Hochul ‘not trying to set up barriers’ for AI companies in New York

New York could have some of the strictest regulations for artificial intelligence models on the books by the end of the year.

That will be up to Gov. Kathy Hochul, who will have to consider the Responsible AI Safety & Education (RAISE) Act by the end of the year. It passed both chambers of the Legislature in June.

The bill was partly in response to a report from artificial intelligence experts and representatives from dozens of countries on the status of emerging models.

One startling finding: it warned that AI is advancing so quickly that it could soon be used to produce biological weapons.

The intent of the legislation is to force artificial intelligence companies doing business in New York to take steps that could prevent that — and other critical events.

It would require developers of frontier AI models to have a safety plan for preventing their product from being used for mass harm. That’s defined as the death or serious injury of at least 100 people or an action that results in at least $1 billion in damages.

The bill would also require developers to report “safety incidents” to the state within three days of them happening, including when the model acts autonomously or is used for an unlawful purpose.

Companies found to have violated the law would face up to $10 million in civil penalties for a first offense and $30 million for incidents after that. 

     

Those are the basics of the bill but it’s far more detailed than I described it. As you may have assumed, it’s faced strong opposition from tech and business groups.

Hochul has tried to champion advancements in AI since she came into office, including the creation of Empire AI at the University at Buffalo, where an extremely powerful supercomputer is being built.

She hasn’t said where she stands on the RAISE Act, which isn’t unusual. Hochul typically doesn’t show her cards before she’s made a decision on a bill, though lawmakers will know if she wants to negotiate changes through a chapter amendment.

That’s when the governor signs the bill with a promise from the Legislature to amend it at the next earliest opportunity.

When asked Tuesday if she wants New York to have some of the country’s strictest regulations of AI, Hochul said she’s instead seeking a balance on the RAISE Act.

“I do support sensible regulations but not to put us at a disadvantage as a state,” she said. “So that’s the delicate nuances I’ll have to be looking at with this legislation.”

Empire AI is a huge effort that’s drawn more than $500 million in public and private funding, for example. Most of that has come from the state but other institutions have also pitched in, like Cornell University and Rensselaer Polytechnic Institute.

That funding has already led to some really interesting projects, like how researchers at the University at Albany are using the supercomputer to try to find a treatment for conditions like ALS and muscular dystrophy.

That’s the public-private side of using AI. Hochul also doesn’t want private investments to run dry in New York.

“I’ll look closely at the legislation but I’m not trying to set up barriers to make our state a place that people do not want to continue to make these investments,” Hochul said.

She has until the end of the year to consider the bill, though could pocket veto it after that if she really wanted to.

————————

TAKE ACTION

(National Council of Nonprofits, 8/20)

Note from Lauri:  Take Action Today – even if you do not currently have staff who take advantage of this Loan Forgiveness program, we MUST send a message to the Trump Administration that we will not stand for these attacks on the nonprofit sector and our ability to recruit and retain the workforce we so desperately need.

Department of Education Proposes Unlawful Overhaul of Public Service Loan Forgiveness

The U.S. Department of Education proposed a new rule on August 18 to unlawfully limit which charitable nonprofits qualify as eligible employers under the Public Service Loan Forgiveness program.

Nonprofit organizations and the public now have 30 days to submit public comments opposing efforts to politicize the PSLF program and make it more difficult for nonprofits to attract the best, most effective workforce needed to serve their local communities.

If implemented, the proposed rule change would limit the types of charitable nonprofits that can serve as qualifying employers – an unlawful change to federal law that extends the program to all charitable nonprofit organizations, regardless of their missions or the communities they serve. The proposed rule opens the door to this and future Administrations changing eligibility for the program based on their priorities or ideology. Without assurances that program rules will remain consistent over the long term, nonprofit professionals cannot rely on the program to make important decisions about their career path. Ultimately, the proposed rule will harm the people and communities that rely on nonprofit organizations and the essential services they provide by making it harder to attract an effective, skilled workforce. For more information on the proposed rule, read NCN’s Insights and Analysis.

Submit Public Comments Today!

NCN urges all nonprofits and advocates to use our comment guide (PDF or Word) to submit public comments by the September 17 deadline opposing the proposed overhaul of PSLF.

Download the Comment Guide
 
 

The comment guide includes everything you need to make your voice heard, including clear directions on how to submit your comments, points to raise in your comments, and background information on the proposed rule.

Thank you for your advocacy!

———————
As many NYS Council members know, The Step Two Policy Project is a think tank that focuses on NYS healthcare policy.  Led by former NYS Budget Director Paul Francis who previously served in several key state leadership roles under several NYS Governors, the Project also features the work of several additional former state government leaders.  Today the Project released the following information, fyi.  Note:  The Tool is primarily focused on NYC MH Programs and Services at this point.  You can follow the The Step Two Policy Project by subscribing to it on the substack app.  
 

By Adrienne Anderson

Step Two Policy Project

Aug 20

Challenge

It is increasingly difficult to differentiate the various programs available for people with mental illness and other behavioral health needs, especially for those who are homeless.

States and localities vary in how regularly they communicate to the public their efforts to pilot, scale, phase out, rebrand, and reconfigure such programs.

     

Too often, the only source of detailed, public information on a given program is an original procurement document designed for those applying to provide the service, or a press release announcing a new initiative. In the latter case, there may be mention of a new program as a brief bullet item beneath the description of a larger budget initiative, and it may be unclear when the program will begin, where it will operate, who will oversee it, and how it will differ from similar programs.

Solution

There are a number of excellent resources that categorize the purpose and operational details of various mental health initiatives. Examples of this include the New York State Office of Mental Health program definitions webpage and a New York City government webpage chronicling mental health efforts going back more than ten years. There are also many examples of recent journalistic stories that demonstrate due diligence in researching the mechanics and funding of new city-, State-, and jointly-funded programs. Yet there is little in the way of a compendium of these services and programs.

I compiled a list of dozens of services into a spreadsheet, initially focused on services targeting homeless adults with serious mental illness in New York City, and identified fields (i.e., columns) that would be useful to know about each program (i.e., row).

In preparing The Step Two Policy Project’s “Reforming Care Management for Adults with Behavioral Health Needs in New York Medicaid,” published in January, and “The Law and Practice of Assisted Outpatient Treatment, Involuntary Removal, and Involuntary Admission in New York State,” published in July, I met with subject matter experts who affirmed the concern described in the “Challenge” section, and helped identify some functional categories to group services within the tool. These categories are designed to be more intuitive and self-explanatory than the familiar administrative conventions, which can seem arbitrary or even internally inconsistent.

The product is an interactive, downloadable worksheet that organizes publicly available information about these various programs. I plan to refresh the worksheet regularly to include suggestions submitted on our website and new information published by government agencies and in news media.

Purposes

  • Provide an interactive, searchable reference for those working in:

    • healthcare and human services, including in case management

    • government, as staff determine which programs to introduce, amend, or close, and where to do so

    • journalism and media, to support their coverage of behavioral health topics

  • Reiterate the value of agencies’ accurate communication of recent information (including performance-related data) about various programs:

    • helps people in need identify programs for which they may be appropriate candidates, and waste less time finding the right resources

    • helps advocacy stakeholders identify programs worthy of promotion and prioritization

Current Limitations

  • Although we would eventually like to expand the services included in the tool, currently, it is not exhaustive.

  • Where known, the tool indicates whether a program/service is also present in non-NYC regions, but the exercise began with a focus on NYC.

    • Still, there are very few program types that exist only outside of NYC (i.e., and not within NYC), so the current approach should be more comprehensive.

  • The initial exercise was focused on services for adults, so there is limited coverage of services focused on age groups that include: children, youth, adolescents, transitional age youth, and alternative names for those groups.

    • Where known, the tool indicates whether a service type is open to youth or families.

  • Eventually, it will be valuable to expand this tool to coverage of a wider range of behavioral health services, but for now, it has limited coverage of substance use disorder (SUD) services, besides services that overlap with services targeting individuals with co-occurring serious mental illness.

  • Beyond Behavioral Health Home and Community Based Services/Community Oriented Recovery and Empowerment (BH HCBS/CORE), there is no discrete coverage of vocational services.

  • Homeless services are currently only captured to the extent that they target individuals with known mental health needs.

User Features

I initially built the tool in Microsoft Excel, and it works similarly to a traditional spreadsheet there. It is hosted on our website using AirTable so that it is fully functional within our Wix website environment. Anyone can utilize the sorting, grouping, and filtering features at the top of the tool. You may also wish to download a CSV file of the current version using the ellipses icon in the top banner menu of the AirTable.

Finally, please submit a suggestion if you have ideas for how to make the tool more useful, or if you would like to share information that should be updated within the tool.

————————————–

Worth the read…

That’s American health care, baby 

STAT, 8/20/25

The effects of a health care system driven by profit have become clear to me in a series of revelations. First, I learned that there exist cures for the rare, fatal “bubble boy” disease, but they are unavailable because drugmakers don’t think they can profit from them. 

Later, I found out that nursing homes, many of which have a track record of providing suboptimal care for our elders, can hide up to two-thirds of their profits in payments shunted to “landlords” and “management companies” that are actually just a nursing home’s owner under a different name.

A third realization came at the end of last year, when we at STAT laid off several employees. One reason we were given? We’d been unexpectedly hit with huge health insurance costs for 2025 — unsustainable ones. 

As our staff processed the news during a Zoom meeting, the realities of working for a self-insured company hit like a ton of bricks: “Wait, so if someone gets cancer, we’re paying for it?” someone asked in the chat. 

Yes, that’s how self-funded insurance plans work. These plans cover about 97 million Americans, which likely includes you, if you work at a company with more than 200 people. Despite the fact that there’s a brand-name insurer on your insurance card, your health care costs are borne solely by your company. And when the price of health care goes up, or more people get care, more of the employer’s money has to go toward those costs, which means it has to make up that money somewhere. 

In this case, that money was a handful of my colleagues.

What does this have to do with AI?

Well, in my new story, I outline an uncomfortable outcome of AI scribes and their closely related siblings, AI billing and coding tools: They are going to drive up the cost of health care. 

Now, I’m not blaming these AI tools for all of the thousands of papercuts that cause health care prices to rise year after year. Everyone in the health care ecosystem has some part to play in that. And it’s possible the net effect of AI products will be to lower the cost of health care, though positive impacts like lowering costs and increased quality may take years to be realized.

But something that is real, quantifiable, and coming soon to a health system near you: AI tools that increase the bills that health care providers charge, based on the more-detailed documentation that ambient AI scribes provide. 

In fact, health tech darlings like Ambience Healthcare and Abridge are actively marketing how their tools can be used to increase billing because they need to show health systems cold, hard American dollars to prove that they’re worth contracting with. (I don’t have room to rehash my entire story here; please go read it!)

Why have you not heard about the possibility of these companies driving up your health care bill before? “I suspect that may be not an angle that the companies think is advantageous for them to widely advertise,” given that they’re riding a huge wave of positive publicity due to the scribes’ ability to reduce doctor burnout, said Caroline Pearson, executive director at the Peterson Health Technology Institute. She told me that it’s “funny” how little data she’s seen on exactly how much money these AI coding and billing tools are returning to health systems, given that the data certainly does exist. 

The person who framed this whole situation the best was Luke Prettol, a principal benefits strategy consultant at AT&T. He helps oversee the insurance benefits the company offers its employees, and has a vested interest in keeping the company’s employees as healthy as possible while also keeping costs as low as possible. He’s really nervous about how these AI tools will affect AT&T’s costs, but at the same time, says that he would adopt these tools if he were a provider.

What we’re seeing play out is that AI is a tool, and different entities — insurers, health care providers, technology developers — are going to leverage it in ways that maximize their interests, said Prettol. “I don’t blame them for doing this. It just makes sense if that’s how you get paid.”

AI has the potential to help people who can’t reach a doctor, help patients be more informed about their care, help doctors catch mistakes, and offload repetitive tasks. But in a profit-driven health care system, we’re going to see the applications of AI that help people make money rise to the top quicker than the others. 

While health care is a business and businesses have to do what’s good for business, the human parts of us may feel queasy about this. And navigating that tension? That’s American health care, baby.