Latest news with #RighttoComputeAct
Yahoo
01-03-2025
- Business
- Yahoo
Lawmakers look to build artificial intelligence framework for Montana
The Montana Senate is seen during the Wednesday, February 12, 2025 session. (Nathaniel Bailey for the Daily Montanan) Montana lawmakers are trying to use this session to build a framework for the usage of artificial intelligence in the state. Lawmakers discussed two AI bills on Friday. Senate Bill 212, brought by Sen. Daniel Zolnikov, R-Billings, would establish the Right to Compute Act, which seeks to protect the individual use AI, and allow the state to restrict it in limited circumstances. It received unanimous approval on second reading, putting it one vote from Senate passage. The Montana Legislature is also taking a close look at the use of AI by health insurance companies, as Rep. Jill Cohenour, D-Helena had a hearing in the House Business and Labor Committee for House Bill 556. Cohenour has several other AI bills, including HB 514, which deals with name, image and likeness rights. The bill passed the Senate on a 97-1 vote on Friday afternoon. Cohenour also has House Bill 513, which would expand and create penalties for use of real and digitally fabricated sexually explicit images. Zolnikov is co-sponsoring all three of Cohenour's House bills. Zolnikov said SB 212 bill provides a framework for using emerging artificial intelligence technology in the state. The bill's intent is not to fully ban artificial intelligence, and Zolnikov said he wants Montana to avoid 'falling in the trap of supporting big tech versus open (source).' It's an attempt to create tech legislation that seeks to avoid freezing out small innovators and anyone else that wants to use artificial intelligence for their own purposes. And even then, there's nothing in the bill explicitly prohibiting or even discouraging a larger company that uses artificial intelligence from setting up shop in Montana, or even the state itself using it . Still, with anything new, problems can arise, and Zolnikov's bill keeps the state's ability to create regulatory language as time passes and the issue evolves. Cohenour's bills address some of these exact situations. 'There's a new world,' Zolnikov said in an interview. 'Let's open the door and then start restricting in a narrow, detailed way, not like other states that are basically trying to ban everything.' Zolnikov's bill preserves the 'right to compute,' with the bill text stating, 'Any restrictions placed by the government on the ability to privately own or make use of computational resources for lawful purposes must be limited to those demonstrably necessary and narrowly tailored to fulfill a compelling government interest.' Additionally, the bill would make any 'critical infrastructure facility' that uses artificial intelligence to run its systems have a risk management policy to fall back on if there are issues. Critical infrastructure facility is a defined term that includes things like electric substations, wastewater treatment plants, dams and oil refineries. Cohenour's AI bills fit into the framework of Zolnikov's bill, addressing some of the 'narrow' issues legislators are looking to fix. On Friday, House Bill 556, which would legislate the use of artificial intelligence in healthcare decisions, got a hearing in the House Business and Labor Committee. 'I've interacted with a couple of the insurance folks, and I asked for a little bit of information from them, and they're all running away from me now,' Cohenour said in an interview Thursday. 'So this is going to be really fascinating tomorrow. I'm sure I'm going to get dog piled on, but really, AI and algorithms should not be determining what kind of healthcare somebody gets. And that's what that bill is all about.' Cohenour's prediction was accurate. BlueCross BlueShield Montana came out strongly in opposition to the bill. The insurance company argued it was not using artificial intelligence to deny medical care, only to approve it. 'BCBSMT may use automated approval technology to approve utilization review requests but never to deny,' BlueCross BlueShield Montana spokesperson Bryan Campen wrote in an emailed response to the Daily Montanan. 'Requests not automatically approved are reviewed by a clinician.' Following the shooting of UnitedHealthcare CEO Brian Thompson, there was a hurricane of anger directed at health insurance providers by the public. One source of that emotion has been artificial intelligence denying claims, and Boston's NPR affiliate reported on a lawsuit filed against UnitedHealthcare alleging exactly that. Class action lawsuits have also been filed against Humana and Cigna for using artificial intelligence to deny insurance claims. While it doesn't appear BlueCross BlueShield Montana has been sued over the issue, Blue Shield of California has. A lawsuit filed last year claimed Blue Shield of California was using a 'unlawful and fraudulent Artificial-Intelligence ('AI') powered prior authorization and review system of their insureds' insurance claims.' Drew Cziok, the director of government relations for BlueCross BlueShield of Montana, spoke against Cohenour's bill, saying it was 'vague.' 'For us, AI augments human judgment,' Cziok said. 'It helps approve services faster. So sometimes this means as soon as everything is submitted, an automated system is able to review a claim and approve it on the spot.' The bill also makes AI software available for inspection by regulatory agencies. 'They should have procedures. They should open their books to make sure we aren't introducing bias when algorithms learn on their own,' Cohenour said on Friday morning. 'That is scary stuff.' The Committee did not take executive action on the bill on Friday. Cohenour's other two bills also deal with some of the narrow issues Zolnikov's legislation allows the state to regulate. House Bill 514, dealing with a person's name, image and likeness, was broadly popular in the House, passing the chamber with just one vote against it. NIL has been most commonly associated with college athletics. Courts have allowed for college athletes to be paid, saying they own their name, image and likeness and should be able to profit off of it. Cohenour's House Bill 514 would extend that legal protection to all Montanans. 'A lot of states are starting to bring this stuff forward, and we all feel like we need to get in front of it,' Cohenour said. 'I mean, it's developing so fast that we're way behind the eight ball.' The Helena representative's other bill strengthens the state's Privacy In Communications laws. It defines 'digitally fabricated,' which is stated in the proposed bill as, 'using technical means, such as artificial intelligence, to create media that realistically misrepresents an identifiable individual as engaging in conduct in which the identifiable individual did not engage.' Breaking the law would be a misdemeanor on first offense, punishable by a $500 fine, six months in county jail, or both. A subsequent offense would be a felony. It also seeks to protect against 'sextortion' which the Federal Bureau of Investigation says it has seen a rise in, specifically amongst minors. This is something Cohenour is concerned about. It has broad bipartisan support and Democratic Minority Leader Rep. Katie Sullivan, D-Missoula, House speaker Pro Tem Rep. Katie Zolnikov, R-Billings, and Senate President Pro Tempore Sen. Kenneth Bogner, R-Miles City, are all co-sponsors. 'Basically the (updates to the) privacy and communication law allows for you to bring an action against somebody that might do a real picture or an AI generated picture, to be able to say that that's also a crime,' Cohenour said. Other bills, including Senate Bill 452 brought by Sen. Daniel Emrich, R-Great Falls, would require, 'disclosures of the use of artificial intelligence by manufacturers of online media.' That bill had a committee hearing on Saturday morning. Additionally, Sen. Janet Ellis, D-Helena, introduced Senate Bill 25, which would outlaw the usage of AI by candidates or political parties within 60 days of an election, unless the media is clearly marked as generated by an artificial intelligence program. That bill passed the Senate's State Administration committee on an 8-1 vote. Sen. Laura Smith, D-Helena, also had legislation move forward to outlaw so-called 'AI deep fakes,' which the proposed bill, Senate Bill 413, refers to as 'synthetic media.' SB 413 moved out of the Senate Judiciary Committee in a unanimous vote.
Yahoo
21-02-2025
- Business
- Yahoo
Opinion: What the federal government can learn from the Mountain West in AI regulation
With artificial intelligence development accelerating, states have been racing to fill a perceived regulatory void. Without a federal framework to show the way, many states have defaulted to the European Union's AI Act as a blueprint for risk-averse AI governance. That heavy-handed approach would be a mistake. As the federal government decides whether to draft a preemptive framework, it should look to the Mountain West as an example. Two states have proposed policies that put consumers in control of their AI-powered futures. Utah passed its AI Policy Act in March 2024, and Montana is in the preliminary stages of its Right to Compute Act. Both approaches focus on providing greater freedoms and benefits to consumers than more precautionary methods in other states. The Utah AI Policy Act is simple. First, it establishes that businesses using AI in their operations cannot blame AI for causing consumer harm. Second, Utah established a regulatory sandbox where developers can roll out new AI products in a controlled environment. Developers are regulated but sheltered from ambiguities in the law while they stress-test their products before full deployment. These two provisions synergize well and give space to innovate while protecting consumers from harm. Prosecuting bad-acting businesses — not AI or its developers — for real harm done to consumers is a better approach to risk mitigation. This precedent already exists in law and prevents using any tool as a scapegoat. On one hand, developers can feel secure knowing they won't be blamed when bad actors misuse their product. Meanwhile, businesses that want to stay in the public's good graces and out of trouble with the law will manage their use of AI to ensure accurate information and compliance with other laws already on the books. This promotes risk reduction without discouraging the use of AI and without adding burdensome compliance costs. By keeping those costs low, younger startups can stand on more equal ground against larger incumbents. The dual benefits of better competition and business accountability set the stage for better quality products and a wider variety of options offered to consumers. Montana has an even simpler approach. Its proposed law recognizes that access to and use of technology is necessary for full participation in society, so it seeks to guarantee the 'Right to Compute' for all Montana citizens. As for risk, where other states would mandate burdensome reports at every level of deployment — Colorado, for example, requires developers and businesses report to the attorney general and consumers — Montana requires deployers to perform annual tests only where AI controls critical infrastructure. This pragmatic approach to risk mitigation allows for less paperwork, freeing up time and resources to spend elsewhere. Montana's proposed legislation puts the technology and decision-making responsibilities firmly in the hands of the people and gives consumers the freedom to enjoy AI sooner. Policies in other states bury the benefits of AI under mounds of requirements and reports, taking up resources and slowing down AI deployment. Such an approach focuses on blocking all the myriad risks while Montana only worries about real harm. Together, passed and proposed legislation in Utah and Montana should serve as a model for the federal government to emulate when drafting their preemptive framework. The consumer-first approach offers greater freedoms and benefits to consumers than other states' risk-focused policies. The Mountain West approach guarantees every consumer the ability to use emergent tech however they like, and sets up incentives in a way that businesses will guard against risk without the need for burdensome compliance costs that would crush young startups. That means more competition and more consumer choice. Together, these laws stand out from their peers. They evoke confidence in Americans to innovate responsibly while maximizing consumer choice and discretion. They do not stifle innovation by preemptively blocking risks. Should the federal government proceed in drafting a preemptive framework, it should incorporate the fundamentally American ideas Utah and Montana are pioneering.
