Only one piece of legislation related to artificial intelligence became law in New Mexico after this year’s legislative session: HB 182 which requires campaigns to provide a disclaimer on any  advertisements or other campaign materials to disclose if artificial intelligence or deep fakes were used to make them.

This was a small step into a larger issue, University of New Mexico experts told the Interim Legislative Science, Technology and Telecommunications Committee Monday what is coming and offered suggestions for the upcoming legislative session.

“It solved one problem, potentially, it at least took a nibble at solving the misinformation and elections problem,” UNM Computer Science professor Melanie Moses said. “I think that one important thing is to take the next step on that legislation and at least consider, do we want to regulate or have laws that consider misinformation more generally right that is harmful, for example, to students or to anyone right, who might have deep fakes of themselves produced and propagated online.”

The committee discussed steps forward in artificial intelligence literacy and ethics.

“I’ve recently realized the extent to which we are sort of standing on the shoulders of giants with this kind of technology, it is integrating what humans have figured out about the world, and then it can make predictions about where we go next. And that is just an incredibly powerful tool,” Moses said.

Story continues below.

AI has been described as just another tool, but Moses thinks it is more than that.

“This is, I think, a form as it’s evolving, a kind of a tool that we haven’t seen before and this is really an opportunity for us to… take control of this, take at least the ability to shape what direction we want it to go in for the citizens of our state, and to contribute to what this is doing for the world at large.”

The meeting included a discussion about AI literacy, during which University of New Mexico College of University Libraries and Learning Sciences Dean Leo Lo described AI literacy “as the ability to understand, engage with, and critically assess artificial intelligence technologies and their societal, ethical, and practical implications.”

The key components of AI literacy are technical knowledge, ethical awareness, critical thinking, practical skills and societal impact, Lo said.

“It feels like the laws and regulations need to catch up right now, just, let’s just say we’re in a gray area. Eventually, hopefully we can move forward, and when regulations have to catch up to this really fast moving technology,” Lo said.

Lo refers to AI as the fourth industrial revolution following the 18th and 19th century steam-based industrial revolution, the late 19th to early 20th century industrial revolution when electricity and mass production came into use and the mid-20th century automation revolution.

Related: Legislators hear talk on AI transparency measures

“The fourth industrial revolution is driven by smart technologies like AI, robotics (and) the Internet of Things. And this revolution is transforming industries by making systems more connected, autonomous and efficient,” Lo said. “So it’s not just automation anymore. It’s about a fully connected, data driven world that requires new skill sets and generative AI, like ChatGPT, like other things like Midjourney, goes beyond simply automating tasks. It creates new content, whether it’s writing images or even solutions to problems, and this technology is transforming fields by assisting with tasks that require creativity and critical thinking, opening up new possibilities in areas like education, healthcare, entertainment.”

AI biases

There are broader implications with AI usage in mainstream life such as inherent biases in AI systems, ethical concerns and potential impacts on job elimination and creation.

One of AI’s biases is racial including mistakes that could affect the rest of a person’s life even after the mistake was fixed.

Such was the case of Michigan man Robert Julian-Borchak Williams who was wrongfully arrested in 2020 due to faulty facial recognition algorithm.

Borchak-Williams was shown security camera stills of a heavy set Black man who did not look like him.

The case was settled in federal court Sept. 28, 2024.

“The implications of this are that arrest records and things like that, the mistakes that deployed software doing facial recognition, has real impact on real people, including that all of the known, publicized cases of false arrests from facial recognition are from African American individuals who, with any sort of human reasoning, would never have been arrested in the first place,” Moses said. “Again, it’s this software- this facial recognition software- is making different kinds of mistakes than humans would make, and humans tend to trust these machines. We tend to think that they are validated, perhaps more than they are, that they are objective in ways that they aren’t and so that’s one of the big challenges. I use this example because it’s so obvious, but it’s certainly not isolated.”

Another case Moses related was about the Black Massachusetts Institute of Technology student who was not recognized by facial recognition until she put on a white mask.

“So it really has this ability to sort of solidify bias that we might have had in the past and if we don’t realize that it’s there, it gives us this veneer of objectivity, right? This, sort of, we think it’s objective, but in fact, it’s not. And if we rely on it, it really makes that bias persist,” Moses said.

Some of these biases AI has include that all doctors are male or that terrorists are Muslim, Moses said.

“We don’t even know what those sort of biases are that could be baked into the data,” Moses said. “And so it’s really important to diversify the data that are training these future AI and that’s another place where I think, here in New Mexico, we can make some real progress. We can actually proactively try to deal with some of those, to counter some of the bias that’s there, and it’s certainly important to the health of the populations of our state.”

AI, jobs and reliability

“AI has the potential to displace millions of jobs, but at the same time, it will create new roles that require different skill sets. So preparing the workforce for these changes is essential to managing this disruption,” Lo said.

This is where AI literacy comes in as a means of driving economic development.

“I think by developing an AI literate workforce, New Mexico has a chance to attract innovative businesses, drive economic growth and ensure that we remain competitive in a fast changing global economy,” Lo said.

Another issue was AI’s lack of reliability.

“It also requires critical thinking, because we need to evaluate the reliability and biases in the AI generated content. They’re not always accurate, they’re not always truthful, and we need to question the use of AI in key areas like governance, education or justice,” Lo said.

Current AI legislation

New Mexico’s Sen. Martin Heinrich, a Democrat, has been active in introducing AI regulatory legislation.

S. 3696, the Disrupt Explicit Forged Images and Non-Consensual Edits Act of 2024, or DEFIANCE Act, passed the U.S. Senate in July, but has not had a vote in the U.S. House.

Heinrich co-sponsored the bill, which passed the Senate on July 23.

Another bill co-sponsored by Heinrich, S. 4569, the Tools to Address Known Exploitation by Immobilizing Technological Deepfakes on Websites and Networks Act or TAKE IT DOWN Act, and other pieces of legislation “have been advanced out of their respective committees of jurisdiction (including our CREATE AI Act) and are being informally negotiated with their House counterparts,” Heinrich spokeswoman Caty Payette told the NM Political Report.

In September, the United States, the European Union and Great Britain signed the first binding treaty concerning AI.“The treaty provides a legal framework covering the entire lifecycle of AI systems. It promotes AI progress and innovation, while managing the risks it may pose to human rights, democracy and the rule of law. To stand the test of time, it is technology-neutral,” according to a Council of Europe press release.