⚡️ News Flash ⚡️

“Less Hype, More Help: AI That Improves Safety, Productivity, and Care”

Senate Commerce, Science, and Transportation

March 3, 2026 (recording linked here)

HEARING INFORMATION

Witnesses and Written Testimony (Linked):

• Ms. Brittany Ng, Vice President, Siemens Digital Industries Software

• Mr. Demetri Giannikopoulos, Chief Innovation Officer, Rad AI

• Dr. Damion Shelton, Co-Founder and Chairman of the Board, Agility Robotics

• Mr. Mark Muro, Senior Fellow, Brookings Institution

HEARING HIGHLIGHTS

QUICK SUMMARY

• Artificial intelligence was discussed as a tool to augment human work across sectors such as healthcare, manufacturing, and logistics. Witnesses described how AI systems were being used to assist radiologists with diagnostic workflows, improve industrial production planning through digital twins, and support robotics deployments in warehouses and factories. Several participants emphasized that these tools were designed to improve productivity and decision-making rather than fully replace human expertise.

• Workforce impacts and job transformation were a central theme throughout the hearing. Senators raised concerns about potential job displacement, while witnesses pointed to labor shortages in healthcare, manufacturing, and logistics as areas where AI and automation could help fill gaps. The discussion also highlighted the need for workforce training, apprenticeships, and new skill development so workers could adapt to evolving roles alongside AI systems.

• Access to high-quality data and the importance of AI governance were repeatedly emphasized. Witnesses explained that machine-readable datasets, including government-held data, could accelerate development and evaluation of AI tools if shared responsibly. At the same time, speakers highlighted the need for privacy protections, transparency, and validation processes to ensure AI systems operate safely and reliably.

• Industrial and sector-specific AI applications were distinguished from consumer-facing technologies. Witnesses noted that industrial AI systems operate in controlled environments using structured datasets from machines and engineering systems, often supporting safety-critical processes in manufacturing and shipbuilding. This distinction was raised in the context of how policy approaches might affect different categories of AI deployment.

• Several members discussed the importance of research investment, talent development, and regional innovation ecosystems. Witnesses highlighted the role of universities, national laboratories, and regional technology clusters in supporting AI research and adoption. The discussion also noted that broader AI literacy and training across the workforce could help expand the benefits of the technology and support its adoption in different sectors and communities.

IN THEIR WORDS

SUMMARY OF OPENING STATEMENTS

• Subcommittee Chair Senator Budd stated that artificial intelligence had the potential to significantly improve quality of life, strengthen economic growth, and enhance U.S. competitiveness. He argued that AI could increase workplace productivity and safety, improve manufacturing capabilities, and support critical industries such as semiconductors and the defense industrial base. He also highlighted AI’s potential to transform healthcare by improving early disease detection, assisting physicians with diagnostics and treatment decisions, and accelerating drug discovery through AI-enabled research tools. Budd acknowledged that many Americans were concerned about the impact of AI on jobs and communities but argued that technological innovation had historically expanded economic opportunity in the United States. He emphasized that maintaining U.S. leadership in AI development and deployment, particularly in competition with China, was essential for national and economic security. He also raised concerns that U.S. businesses might lag international competitors in adopting AI technologies and stressed the importance of scaling AI across the economy and manufacturing sector to strengthen domestic production.

• Subcommittee Ranking Member Senator Baldwin emphasized that artificial intelligence could improve worker safety and productivity but raised concerns about the risks associated with deploying AI systems without adequate safeguards. She highlighted reports that federal officials had pressured AI companies to deploy technologies without appropriate guardrails and cited the case of Anthropic being labeled a security risk after refusing to support certain applications of its technology. Baldwin noted that many workers were concerned about the reliability and accuracy of AI systems, potential job displacement, and the lack of human oversight. She also pointed to broader concerns related to privacy, intellectual property, and the energy demands of AI systems. Baldwin argued that AI could improve workplace safety, reduce burnout, and increase productivity if workers were involved in the design and deployment of the technology. She emphasized that workers performing day-to-day tasks often had the best understanding of workplace inefficiencies and safety risks and should have a role in shaping how AI systems are implemented.

SUMMARY OF WITNESS STATEMENTS

• Mr. Giannikopoulos stated that artificial intelligence was being used to assist clinicians in diagnosing medical conditions more quickly and accurately. He noted that diagnostic errors contribute to a significant number of deaths and disabilities each year and said these errors can occur when clinicians face high patient volumes and increasingly complex medical information. Mr. Giannikopoulos explained that AI tools can analyze medical images, highlight potential findings for radiologists, integrate clinical guidance into physician workflows, and improve communication among care teams. He said these tools may help accelerate treatment decisions in time-sensitive cases such as aortic dissection and stroke and may assist in ensuring patients receive appropriate follow-up care. Mr. Giannikopoulos stated that AI systems operate alongside physicians and within existing healthcare regulatory frameworks, and he emphasized the importance of governance, monitoring, and consistent standards for AI tools after deployment, particularly in healthcare settings with limited access to specialized expertise.

• Ms. Ng discussed the use of artificial intelligence in industrial and manufacturing settings. She described how industrial AI, digital twins, and software-based automation are being used to connect design, simulation, and production processes in industries such as shipbuilding and manufacturing. Ng said these technologies can help manufacturers identify production bottlenecks, improve quality control, reduce machine downtime, and manage energy use. She explained that digital shipbuilding platforms allow companies to simulate construction processes in virtual environments before physical production begins. Ng also noted that many manufacturers face workforce shortages and said AI tools may assist workers by reducing repetitive tasks and supporting more complex production environments. She suggested that policymakers consider the differences between industrial AI and consumer applications and highlighted the role of federal programs and partnerships in supporting technology deployment.

• Dr. Shelton discussed the development and deployment of humanoid robots designed to assist with repetitive tasks in warehouses and manufacturing environments. He stated that technological advances have historically increased productivity and influenced labor patterns, citing changes in agricultural employment over the past two centuries. Dr. Shelton described Agility Robotics’ humanoid robot, which is used by companies in logistics and manufacturing to perform material-handling tasks alongside human workers. He noted that some industries are experiencing labor shortages and high turnover in physically demanding jobs. Dr. Shelton said robotics and AI technologies can shorten the time required to develop new applications for automated systems. He also stated that safety considerations remain important as robotics technologies develop and emphasized the need for testing and standards in controlled environments. Dr. Shelton additionally noted that China has made rapid progress in humanoid robotics development and manufacturing.

• Mr. Muro stated that recent developments in artificial intelligence have introduced new tools that expand the capabilities of individuals, firms, and entrepreneurs. He noted that discussions about AI include both optimism about its potential benefits and concerns about adoption challenges, safety, and workforce impacts. Mr. Muro said continued federal support for research and development could help maintain U.S. leadership in AI technologies. He discussed proposals to increase investment in AI research, expand collaborative testing environments involving public and private institutions, and support applications of AI that assist workers and decision-making processes. Mr. Muro also referenced proposals to encourage regional technology development through initiatives such as innovation hubs and regional investment zones. He stated that workforce readiness will depend on both the development of specialized AI talent and broader efforts to improve AI literacy and skills across the workforce.

SUMMARY OF KEY Q&A

• Sen. Budd asked how expanded access to federal machine-readable data would affect companies’ ability to deploy AI, particularly in light of the Open Government Data Act’s requirements. Mr. Giannikopoulos said broader access to outcomes-linked validation data would help healthcare AI developers evaluate performance against real-world conditions and support adoption.
  Sen. Budd asked what specific radiology datasets would be most useful and how they could be shared without violating privacy. Mr. Giannikopoulos said access to both images and reports, linked to longitudinal outcomes and appropriately de-identified, would enable better development and assessment of tools that track progression and follow-up. Ms. Ng said trusted, usable data was central to industrial AI and that better access to machine, engineering, and manufacturing-system data improved the quality of AI insights in production environments. Dr. Shelton said robotics firms had limited ways to obtain training data and that government datasets could help identify high-value use cases and support broader understanding of where robots could improve work.
  Sen. Budd asked how census-related information could translate into actionable opportunities for robotics deployments. Dr. Shelton said integrating labor, injury, and performance data could help target tasks that improved safety and health and better align automation offerings with workforce needs.
  Sen. Budd asked whether documenting labor shortages could reduce fears about job replacement and improve adoption of robotics. Dr. Shelton said it could, and he added that smaller businesses had been interested in automation but had lacked feasible solutions earlier in the decade.
  Sen. Budd asked how robotics capabilities in 2026 compared with what was possible in 2020. Dr. Shelton said capabilities had advanced substantially and that recent AI improvements had lowered barriers to serving smaller and mid-sized businesses, though rollout choices still mattered.

• Sen. Baldwin asked where AI offered the most promising opportunities to expand domestic shipbuilding capacity, how Siemens worked with labor and workers, and whether job-displacement fears could limit workforce growth. Ms. Ng said AI could improve shipbuilding across design, production, and sustainment by accelerating design iteration, optimizing sequencing and material flow, and supporting predictive maintenance, and she added that digitally native workers were increasingly attracted to modern tools.
  Sen. Baldwin asked what local officials should seek when negotiating with data center developers to ensure long-term local benefits. Mr. Muro said communities could negotiate early for partnerships tied to regional research, workforce, and energy innovation, including potential trades involving faster permitting in exchange for deeper local investment.

• Sen. Cruz asked how witnesses responded to public fears that AI would reduce employment and take jobs. Mr. Giannikopoulos said healthcare already faced major clinician shortages and rising imaging demand, and he said AI was intended to reduce workload pressures while increasing the capacity of clinicians rather than replacing them. Ms. Ng said industrial AI tended to expand demand for skilled trades and that the larger risk to jobs was loss of industrial competitiveness rather than modernization. Dr. Shelton said automation helped address labor gaps in logistics and also changed job content over time, and he said the economy historically adapted as technologies reshaped roles.
  Sen. Cruz asked what witnesses hoped AI would enable that was not yet possible and what AI use had surprised them. Mr. Giannikopoulos said AI could move healthcare further toward earlier, more personalized, and predictive pathways that shortened diagnostic delays and improved prevention. Ms. Ng said Siemens sought lifecycle-wide adoption rather than siloed use and said she was surprised by the Navy’s application of industrial AI to model shipyard capacity and infrastructure planning. Dr. Shelton said he was surprised that an early version of ChatGPT could generate working robot-control code on the first attempt and said this pointed toward faster deployment from natural-language task descriptions.

• Sen. Blunt Rochester asked how companies engaged workers when designing and implementing AI changes and how Congress could help ensure workers were trained and not left behind amid disruption. Ms. Ng said Siemens found adoption worked best when workers were centered and pointed to shipyard leadership examples where frontline pain points informed technology deployment. Mr. Muro said projections about disruption were directional and that impacts could be beneficial or harmful, with outcomes depending on how transitions were managed.
  Sen. Blunt Rochester said she wanted specific strategies to include workers, support reskilling, and prevent harmful disruption as AI changed tasks and reduced barriers like coding.

• Sen. Blackburn said Congress needed guardrails for AI and asked how innovators should approach privacy in a digital health ecosystem. Mr. Giannikopoulos said healthcare AI deployments operated under HIPAA and auditable electronic access controls, and he said institutions were treating AI as part of broader workforce transformation to deliver new care pathways.
  Sen. Blackburn asked Ms. Ng to distinguish industrial AI from consumer AI and explain why the difference mattered for policy. Ms. Ng said industrial AI relied on controlled, safety-critical datasets from machines and engineering systems and that humans remained the decision authority for actions based on AI recommendations.

• Sen. Moreno argued that repeated historical predictions of automation-driven job loss had not materialized in overall employment data and asked what was different about the current AI period. Ms. Ng said the key difference was the ability to align AI tools to specific operational outcomes through close partnerships that clarified goals and roadmaps.
  Sen. Moreno asked whether institutions that train workers were prepared for the unusually rapid pace of AI change. Dr. Shelton said the speed was different and that knowledge work was now being affected in new ways, requiring workers to adapt to managing AI-enabled tools, and he suggested education might pivot toward broader skills associated with the liberal arts.

• Sen. Hickenlooper asked how RAD AI tested products before use and whether independent evaluation frameworks could build trust. Mr. Giannikopoulos said RAD AI was built with radiologist-led design, relied on workflow validation with clinicians, and emphasized trust through transparency and integration of trusted clinical resources such as RSNA knowledge at the point of care.
  Sen. Hickenlooper asked how Siemens ensured transparency with customers about AI design limits and enforceable service expectations. Ms. Ng said Siemens supported transparency by showing the source data behind outputs and providing explanations for recommendations, supported by training and oversight roles to interpret systems.

• Sen. Young asked how industrial AI could help restore U.S. shipbuilding dominance and why consumer-focused AI policy should not impede industrial deployment. Ms. Ng said industrial AI served as a force multiplier across design, production, and sustainment by accelerating simulation, improving sequencing, reducing rework, and supporting predictive maintenance to keep equipment available.
  Sen. Young asked how Siemens was upskilling the workforce for AI-enabled industrial roles. Ms. Ng said Siemens was developing micro-credentials with partners and had committed to training 200,000 electricians and manufacturing experts by 2030 through collaborations with education and workforce institutions.

• Sen. Cantwell emphasized AI-accelerated science and asked how AI and automation could compress biology and chemistry lab work from years to months and why scientific workforce capacity mattered. Mr. Muro said AI progress depended on sustained basic science activity and talent because scientific work generated the data and ideas needed for iterative innovation. Dr. Shelton said general-purpose robotics could connect “islands of automation” in labs and that narrow AI systems could achieve outsized performance on specific problems, enabling faster transitions from computation to automated wet-lab testing.
  Sen. Cantwell asked how to explain the value of this acceleration to the public and whether applying AI to existing research investments could yield major gains. Dr. Shelton said AI could help translate human ideas into testable physical implementations faster and shorten the cycle from concept to controlled experimentation.

• Sen. Rosen asked Ms. Ng what trust and reliability standards industrial customers demanded and what NIST’s AI standards efforts should consider. Ms. Ng said trust depended heavily on strong underlying machine, manufacturing-system, and engineering data that formed an authoritative digital backbone for reliable analysis and recommendations.
  Sen. Rosen asked what barriers smaller customers faced in adopting AI and what Congress could do to prevent an adoption gap. Ms. Ng said barriers centered on awareness, ability, and willingness, including understanding use cases, aligning incentives and worker engagement, and managing organizational change for rollout.
  Sen. Rosen asked how AI literacy could close adoption gaps for smaller businesses and communities. Mr. Muro said regional technology and learning ecosystems could support broader adoption and that federal policy could more directly strengthen regional capacity tied to talent development.

• Sen. Baldwin asked Mr. Muro how registered apprenticeships could incorporate AI and support retraining in an AI-impacted labor market. Mr. Muro said apprenticeships could combine hands-on experience with AI-enabled learning supports and could become more important as pathways into specific skills and careers amid changing credential dynamics.

• Sen. Budd asked why predictions that AI would eliminate radiologists had been wrong and what that implied for future AI forecasts. Mr. Giannikopoulos said radiology required human judgment for interpretation and edge cases, that AI was trained on clinician-generated outputs and could not replace clinical synthesis, and that AI was better suited to surfacing information and augmenting care, including in rural systems.

ADD TO THE NIMITZ NETWORK

Know someone else who would enjoy our updates? Feel free to forward them this email and have them subscribe here.