HB 2298 proposes the creation of a new grant program under the Health and Human Services Commission (HHSC) to support the use of artificial intelligence (AI) in medical imaging for cancer detection. While the stated public health goal—to improve early cancer detection and reduce mortality, is laudable, the structure, funding mechanism, and long-term implications of this bill present multiple areas of concern from a limited-government, fiscally conservative perspective. For these reasons, Texas Policy Research recommends that lawmakers vote NO on HB 2298.
HB 2298 directly expands the size and responsibilities of state government by requiring HHSC to establish and administer a new grant program, including adopting rules, overseeing the application process, awarding and managing funds, and evaluating results. This constitutes an expansion of government into the technological and diagnostic domains of healthcare, areas that are traditionally and effectively driven by the private sector. It adds a new administrative mandate and creates a new function within HHSC that does not currently exist.
The bill also grants rulemaking authority to the executive commissioner, effectively opening the door to future regulatory structures around the use of AI in medicine. While this is framed as oversight of grant compliance, it creates a precedent for state-sanctioned guidelines for how private providers implement AI technologies. This could grow into a more intrusive regulatory apparatus over time, contrary to conservative principles of minimal state interference in private clinical or business operations.
The fiscal note for HB 2298 estimates a $2.75 million cost to the General Revenue Fund over the first biennium and recurring costs of $1.37 million annually thereafter. The program authorizes up to five $250,000 grants per year, along with full-time staff and administrative support. While these numbers may seem modest in comparison to the overall state budget, they represent an unnecessary and unjustified taxpayer subsidy for a program that should be market-driven. The bill does not provide a compelling justification for why state funds should underwrite the early adoption of technologies that already have strong commercial interest and investor backing.
Furthermore, there are no embedded performance-based funding contingencies. While recipients must submit a report after one year, there is no requirement for measurable clinical outcomes or cost savings to justify continued funding. This lack of accountability, combined with the fact that the grant program sunsets in 2035, a full decade after its launch, places the financial burden of experimentation on taxpayers without adequate safeguards for return on investment.
The program’s structure creates an uneven playing field by limiting participation to five recipients per year, potentially favoring well-resourced, urban, or politically connected hospitals or health systems. Even with the 10% matching fund requirement, the cost-sharing threshold is low enough that large providers can easily absorb it, while smaller or rural facilities may be effectively shut out of participation. This kind of selective subsidy risks distorting the competitive healthcare marketplace by directing public funds toward favored institutions or early adopters, rather than letting the value of AI tools be determined by market demand and provider choice.
There is also no requirement in the bill for vendor neutrality or competitive bidding processes for the AI technology itself. Without these provisions, the state may unintentionally favor specific companies or products, resulting in a government-endorsed vendor model, a classic example of crony capitalism, where private entities benefit from public-sector privilege rather than market success.
Although HB 2298 does not impose new mandates, it creates a framework that could easily evolve into a regulatory regime. By granting rulemaking authority and requiring compliance with AI oversight guidelines, the bill sets a precedent for future legislation to regulate or standardize how AI is used in clinical settings. Once established, such programs often lead to mission creep, expanding beyond their original intent and increasing government influence over medical decision-making and innovation. Lawmakers should remain vigilant about such entry points into healthcare regulation.
AI-based diagnostic tools are already being developed and adopted in private health systems across the country, funded by venture capital, hospital system investment, and philanthropic grants. If the technology is effective, providers have ample financial and clinical incentive to adopt it without government subsidy. The role of the state is not to underwrite pilot programs for commercial technologies, especially when those technologies are still evolving and carry unresolved questions about efficacy, accuracy, bias, and data privacy. The best approach would allow the free market to determine whether AI-based diagnostics succeed, without government distortion or taxpayer support.
In sum, while the bill is motivated by good intentions, improving cancer screening and outcomes, it is built on a flawed model that increases the size of government, imposes unnecessary costs on taxpayers, risks favoritism and market distortion, and sets a dangerous precedent for regulatory creep into medical technology. These concerns directly contradict core conservative principles of limited government, fiscal restraint, and free market integrity.
- Individual Liberty: The bill does not explicitly restrict individual rights or medical choice, nor does it mandate the use of artificial intelligence (AI) in cancer detection. Participation by healthcare providers is voluntary, and the AI tools would still be used under physician oversight, not as standalone diagnostics. However, the bill lacks provisions to protect patient autonomy, such as requiring informed consent for AI-assisted diagnostics or clarifying how AI-generated medical results will be disclosed to patients. Moreover, there is no clear protection of individual health data privacy in the grant structure. If AI vendors or grant recipients are allowed to use or sell de-identified patient data, it could raise ethical and constitutional questions related to medical consent and privacy, issues directly tied to personal liberty.
- Personal Responsibility: The bill neither enhances nor undermines personal responsibility in a meaningful way. It creates no new obligations for individuals, nor does it incentivize behavioral changes such as increased screening participation. One could argue that early detection tools support individuals in making proactive health choices, but this is an indirect effect, not a policy feature.
- Free Enterprise: The bill creates a state-funded grant program for a specific kind of healthcare technology, thereby disrupting a market that should be driven by competition, demand, and innovation. By awarding up to $250,000 to five applicants annually—with minimal competition or outcome-based criteria, the bill gives a select group of healthcare providers and potentially their tech vendors a taxpayer-funded head start, rather than allowing natural market forces to decide which tools are most effective. Additionally, the bill does not require vendor neutrality or open procurement, meaning public funds could favor specific companies or technologies. This introduces the risk of cronyism, suppresses innovation, and violates the principle that government should not pick winners and losers in the market.
- Private Property Rights: The bill doesn’t directly interfere with ownership of land or physical property. However, there are unresolved questions about ownership of AI-generated diagnostic data. If private patient data or image scans used in AI systems are shared, stored, or monetized by vendors or the state without clear contractual and statutory protections, this could indirectly infringe on patient and provider data ownership rights, a modern extension of the property rights principle.
- Limited Government: This is the area where the bill most clearly violates a core liberty principle. The bill creates a new grant program, expands HHSC’s regulatory authority, grants rulemaking powers to the executive commissioner, and funds both program administration and ongoing grants with state resources. Even though the bill includes a sunset clause (2035), there is no built-in mechanism for automatic program review or defunding if benchmarks are not met. This is a clear case of government expansion, both in budgetary terms and administrative function, and it contradicts the principle that the state should do only what is necessary to protect individual rights, not underwrite private sector innovation.
