FAQ for Texas TRAIGA skeptics: Hard questions answered here

Mural of Willie Nelson in Austin, Texas

Texas artists like Willie Nelson are seeing their intellectual property infringed by out-of-state AI companies. TRAIGA would protect Texans against the harms of AI while establishing a framework that allows local tech companies to thrive.

Feb. 17, 2025 — This month state lawmakers in Austin will take up the Texas Responsible AI Governance Act (TRAIGA), one of the most significant pieces of state-level AI legislation in the country.

The Transparency Coalition has prepared a complete guide to TRAIGA (in plain English), available here.

The FAQ below is meant to answer many of the legitimate questions that may arise over the course of the coming weeks as lawmakers in Austin consider the landmark piece of legislation.

 Q: I’m against heavy-handed regulation. why should i consider traiga? 

TRAIGA is the Texas Responsible AI Governance Act, and responsible may be the most important word in the title.

This is appropriate, balanced, and reasonable governance of an industry and a technology that currently has no governance, no rules, and no fairness for Texans. Artificial intelligence has a profound impact on the lives of all Texans in the same way as insurance, health care, education, and air travel—and those industries thrive under rules that protect both consumers and business.

Q: who’s behind the bill?

TRAIGA was crafted by Rep. Giovanni Capriglione, a conservative Republican lawmaker and a successful small businessman, in consultation with stakeholders all over the state. He knows what it takes to encourage business to thrive in Texas.

Q: Why would business want this bill?

AI developers (the companies that create an AI system) and AI deployers (the companies that use AI systems in their daily work) currently face legal uncertainty over the rules of the road—because there are none—and the potential legal liability they may face if consumers are harmed by their AI use. TRAIGA offers them legal assurance and the certainty they need to invest in ethical, industry-leading products.

Q: how would traiga affect property rights? 

Lawmakers skeptical of regulation are often strong defenders of property rights. Currently, absent any rules whatsoever, California-based AI developers are violating the property rights of Texas authors, songwriters, artists, photographers, and other creative professionals by using their work without permission to train AI systems. This is like stealing another man's truck to perform a day’s work, then returning the truck at suppertime and calling it all good.

By requiring AI developers to be transparent about the legal use of training data, TRAIGA would drive the bad actors out of the market and reward ethical AI developers who partner with intellectual property owners.

Q: Are we moving too soon? Isn’t this based on speculative harm that hasn’t yet occurred?

Algorithmic discrimination and the use of AI in consequential decision making is having real-life impacts on Texans right now. The risks are real. The harm is real.

A recent report in MIT Technology Review revealed the many ways that AI systems are already giving machines control over our health care, our housing, our education and our employment. The term “excoded” is gaining traction as a description of a person—a real human—excluded or harmed by a machine-based decision. The author wrote:  

“You can be excoded when a hospital uses AI for triage and leaves you without care, or uses a clinical algorithm that precludes you from receiving a life-saving organ transplant. You can be excoded when you are denied a loan based on algorithmic decision-making. You can be excoded when your résumé is automatically screened out and you are denied the opportunity to compete for the remaining jobs that are not replaced by AI systems. You can be excoded when a tenant-screening algorithm denies you access to housing. All of these examples are real. No one is immune from being excoded.”  

Q: Should we be creating a patchwork of state laws around AI? Why not wait for a bill from Congress that covers all states?

No matter which party is in charge, Congress has shown little ability to lead or act to protect Americans from the most extreme risks posed by artificial intelligence. Democrats haven’t passed a law. Republicans haven’t passed a law. In Congress, each party is more interested in making sure the other side fails.

State legislators, by contrast, are primarily concerned about protecting their fellow citizens from the harms of AI while encouraging local innovation and economic growth. State-level AI proposals like TRAIGA are carefully crafted to align with similar language and requirements established in other states. This gives ethical AI companies the certainty they need to innovate in ways that comply with all state laws, fosters strong industry standards, and allows them to thrive in the global marketplace.

Finally, national consensus on emerging issues often develops at the state level. If and when Congress acts to pass a national artificial intelligence law, it will undoubtedly be shaped and informed by the laws already crafted and tested at the state level.

Q: I’m concerned that AI training data record-keeping would burden Texas AI developers, giving an advantage to out-of-state companies.

TRAIGA governs not just Texas-based AI companies but any AI company serving Texans. That includes AI developers in other states and other countries.

Knowing the provenance of an AI system’s training data is a basic part of legal vetting and quality control. It allows a local Texas company to know that the AI system it’s investing in, and deploying, has been sourced with high-quality, legally obtained training data. Transparency in AI training data protect Texas developers and deployers from downstream copyright and privacy lawsuits.

Businesses in a myriad of other industries—health care, financial service, transportation, manufacturing—are required to keep records of their work. Record-keeping is a normal part of any business, including the tech industry. Developers and deployers already keep logs of the information required by TRAIGA. The idea that AI developers should be granted a special exemption to dispense with ordinary record-keeping is absurd.

Q: could TRAIGA hinder AI innovation by Texas developers?

The few high-risk use cases that are forbidden under TRAIGA are clearly against the values of Texas and put Texans at great risk of harm. The provisions contained within TRAIGA are hardly onerous. They are appropriate and balanced, and any AI developer with an ethical business model will quickly and easily adapt the TRAIGA’s requirements.

Businesses desire legal certainty and stability. The chaotic legal uncertainty that exists around AI today rewards bad actors and punishes companies striving to create ethical, legally sourced, high-quality AI systems that serve the best interests of individuals and society.

Q: Would TRAIGA constrain Texas start-up companies?  

No. In fact, TRAIGA would create a more level playing field for everyone. Our current no-rules AI environment is concentrating enormous power in a handful of big corporations, most of them based in California. TRAIGA will allow smaller entrants to compete more effectively against these global giants, increasing competition and providing more options to consumers and customer businesses.

Deployers of AI systems are typically smaller companies advancing specific use cases—a durable goods manufacturer using an AI system developed by OpenAI or Meta, for example. Currently, the large tech developers are shifting all potential liability for AI errors or algorithm discrimination onto these smaller deployers. TRAIGA balances liability and provides deployers with the legal certainty they need to thrive without fear of getting sued.

TRAIGA also contains an exemption for small businesses AI start-ups, which allows Texas entrepreneurs to innovate, grow, and come into compliance with Texas’s AI rules when they have sufficient resources.

Finally, TRAIGA establishes an AI Sandbox Program that will give Texas companies the space they need to experiment and innovate in an environment that’s safe for both the company and individual Texans.

Q: We can’t afford to lose the AI race to China.  

The answer to the challenge of China isn’t going to come out of a wholesale surrender to a handful of California corporations. American success in AI will require grit, creativity, ingenuity, and out-of-the-box thinking. Those are things Americans, and especially Texans, have always been pretty good at.

TRAIGA’s requirements will apply to all AI companies, foreign and domestic, who engage with Texas consumers. As we’ve seen with TikTok, American laws can successfully be applied to foreign corporations to make them abide by the rules of the road.  

Texas offers a large base of tech talent and a thriving system of research institutions that will attract even more talent with TRAIGA’s safeguards and legal assurance. The rules established in the bill will remove bad actors and encourage more tech entrepreneurs to create and grow in the fertile ground of Texas.

Q: Aren’t our current anti-discrimination laws sufficient to deal with AI?

You would think so, but unfortunately they aren’t. Current discrimination laws focus on intentional discrimination by individuals or organizations, not digital algorithms. An automated decision-making tool (ADMT) using artificial intelligence can have disparate and sometimes devastating impact without an overt intent to discriminate.

TRAIGA makes clear the steps companies must take to mitigate the harmful effects of AI-enabled algorithmic discrimination. This unfair machine bias can deny jobs, housing, medical procedures, educational opportunities, and basic service access to older Texans, or people living within a rural ZIP Code, or with a certain credit score, or people of a certain sex or ethnic background.

By prohibiting the sale of unfairly biased AI systems and requiring AI developers to fix illegal bias as it arises, TRAIGA protects all Texans from having their life opportunities and freedoms unfairly quashed by a machine.

Q: I’m concerned that TRAIGA’s enforcement fines are too onerous. Or too lenient.

Enforcement fines can be a Goldilocks problem: Not too big, not too little, it’s tough to find the amount that’s just right.

The proposed fines in TRAIGA are significant. They range from $50,000 to $200,000 per violation.

That level of penalty is necessary because the dominant AI corporations right now are Meta, Microsoft, Alphabet (Google), and other Silicon Valley giants. These are companies with annual revenue in the hundreds of billions of dollars. Without a significant fine, they could continue to flout Texas law and simply write it off as a relatively small cost of doing business.

Importantly, TRAIGA contains a “right to cure” clause. This means that companies suspected of violating the AI law would be given proper notice and 30 days to fix the problem. If they choose to ignore the notice and continue to violate the law, then the enforcement fines would kick in.

learn more about ai bills in texas

 

Previous
Previous

Texas AG opens investigation of DeepSeek, testing the power of state law over AI

Next
Next

Two big rulings: Courts are starting to expose AI piracy of copyrighted material