September 15, 2026.
Artificial intelligence is becoming part of everyday business operations while lawmakers in Washington consider different approaches to managing the technology.
For companies, the debate is no longer limited to whether artificial intelligence can improve productivity. Businesses are also examining the quality of the data used by artificial intelligence systems, how employees interact with the technology, whether investments produce measurable results and what responsibilities may eventually accompany artificial intelligence adoption.
A recent analysis by Supply Chain Dive illustrates some of those challenges. The publication reported that 56 percent of chief supply chain officers surveyed by Gartner identified integrating artificial intelligence with older computer systems and business processes as a major hurdle, while 50 percent cited limited internal expertise and talent. (supplychaindive.com)
The experience of supply-chain companies offers a useful business lesson as policymakers consider how artificial intelligence should be governed.
Artificial intelligence adoption begins with the business problem.
Supply Chain Dive reported that companies should look for situations in which decisions occur frequently, under time pressure and with data the organization can actually access.
That can make recurring decisions such as ordering, inventory deployment and delivery routing more promising candidates for artificial intelligence than occasional strategic exercises.
Tractor Supply Company, for example, has been using artificial intelligence to assist with delivery route building.
Data availability is another consideration. Supply chains often depend on suppliers and logistics companies for information. If a business does not have reliable visibility into the information needed for a decision, automating the process will not solve the underlying data problem.
Companies therefore face a preliminary question before investing in artificial intelligence: Is the information necessary to make the decision actually available and reliable?
Pilot projects can expose problems before companies spend more.
Businesses are also learning that an artificial intelligence pilot can fail even when the underlying technology appears promising.
Supply Chain Dive reported that experts recommend establishing the purpose of a pilot, confirming that the necessary data exists, determining whether a return on investment can be demonstrated and making sure employees are comfortable using the technology.
Employee knowledge can be particularly important. Planners, inventory managers and other workers may possess operational knowledge that has never been formally recorded.
That knowledge can help companies determine whether an artificial intelligence model is making reasonable recommendations or missing circumstances that experienced employees recognize immediately.
Knowing when to stop is part of artificial intelligence management.
Starbucks provides one example of an artificial intelligence project that did not continue.
The company abandoned an artificial intelligence inventory-management system after approximately nine months. The system used computer vision to help automate inventory counting. Supply Chain Dive reported that Reuters had found the system sometimes miscounted or mislabeled items.
The lesson for other businesses is not necessarily that artificial intelligence inventory systems do not work. Rather, companies need standards for determining whether an individual application is producing enough value to justify continued investment.
Supply Chain Dive recommends monitoring performance indicators and determining whether a system continues to perform adequately after adjustments have been made.
That puts artificial intelligence in a familiar business category: an investment that should be measured against its results.
House of Representatives
FRONTIER Act focuses on advanced artificial intelligence
The House of Representatives is considering a different question: how the federal government should oversee developers of the most advanced artificial intelligence systems.
House Resolution 9925, the Frontier Risk Oversight, National Transparency, Independent Evaluation, and Reporting Act, or FRONTIER Act, was introduced in the House of Representatives on July 23, 2026.
Representative Jay Obernolte of California introduced the bill with five bipartisan cosponsors: Representative Lori Trahan of Massachusetts, Representative Erin Houchin of Indiana, Representative Scott Peters of California, Representative Scott Franklin of Florida and Representative Suhas Subramanyam of Virginia.
The legislation was referred to the House Committee on Energy and Commerce and, in addition, to the House Committee on Science, Space, and Technology.
The legislation's stated purpose is to establish federal oversight of the development and deployment of frontier artificial intelligence in interstate and foreign commerce.
That is narrower than a general artificial intelligence law covering every company that uses an artificial intelligence application.
The FRONTIER Act is directed toward developers of advanced frontier artificial intelligence systems and would establish requirements involving areas such as risk management, evaluation and reporting.
The legislation remains under consideration and has not become federal law.
For businesses, that distinction is important. A company using an artificial intelligence-powered accounting, marketing, inventory or customer-service application is not automatically subject to the FRONTIER Act simply because it uses artificial intelligence.
The potential direct regulatory impact would be concentrated on the developers and systems covered by the legislation.
United States Senate
Senate proposals take several different approaches
The United States Senate has not produced one comprehensive artificial intelligence regulatory bill that governs every aspect of the technology.
Instead, Senators have introduced separate measures addressing specific issues.
Those proposals range from financial-sector experimentation and workforce planning to artificial intelligence-generated content, competition and federal government use of artificial intelligence.
Financial services: Senate Bill 2528
Senate Bill 2528, the Unleashing Artificial Intelligence Innovation in Financial Services Act, was introduced on July 29, 2025, by Senator Mike Rounds of South Dakota, with Senator Martin Heinrich of New Mexico, Senator Thom Tillis of North Carolina and Senator Andy Kim of New Jersey.
The bill would establish Artificial Intelligence Innovation Labs allowing certain financial-sector participants to experiment with artificial intelligence without an expectation of enforcement actions for qualifying activities.
The measure was referred to the Senate Committee on Banking, Housing, and Urban Affairs.
Rather than treating regulation solely as a restriction, the proposal takes a controlled-experimentation approach.
For financial businesses, that could provide a mechanism for testing new artificial intelligence applications while regulators observe how the technology performs.
Artificial intelligence safety evaluation: Senate Bill 2938
Senate Bill 2938, the Artificial Intelligence Risk Evaluation Act, was introduced on September 29, 2025, by Senator Josh Hawley of Missouri, with Senator Richard Blumenthal of Connecticut.
The bill would require the Secretary of Energy to establish an Advanced Artificial Intelligence Evaluation Program.
This proposal takes a different approach from the financial services bill. Instead of creating a mechanism for businesses to experiment with, it focuses on building federal capacity to evaluate advanced artificial intelligence risks.
The legislation was referred to the Senate Committee on Commerce, Science, and Transportation.
Workforce planning: Senate Bill 3339
Senate Bill 3339, the Artificial Intelligence Workforce PREPARE Act, was introduced on December 3, 2025, by Senator Jim Banks of Indiana, with Senator Maggie Hassan of New Hampshire, Senator John Hickenlooper of Colorado and Senator Jon Husted of Ohio.
The bill would improve federal efforts to measure and forecast artificial intelligence's effect on the American workforce.
The proposal calls for collecting information about artificial intelligence adoption and artificial intelligence-related layoffs, improving occupational forecasts and providing better information for workforce-training programs.
It would also give the Department of Labor authority to recruit temporary experts in artificial intelligence, machine learning and advanced data science.
The legislation goes beyond simply studying artificial intelligence's effect on employment.
Under proposed changes to federal worker-adjustment reporting requirements, certain employers involved in mass layoffs would have to provide information concerning the role artificial intelligence played and identify any efforts made to increase workers' skills or retrain employees before the layoffs.
That could make workforce preparation an important part of how some businesses document artificial intelligence-related changes.
The bill was referred to the Senate Committee on Health, Education, Labor, and Pensions.
Election administration: Senate Bill 2346
Senate Bill 2346, the Preparing Election Administrators for Artificial Intelligence Act, was introduced on July 17, 2025, by Senator Amy Klobuchar of Minnesota, with Senator Susan Collins of Maine and Senator Mark Kelly of Arizona.
The proposal would direct the Election Assistance Commission to develop voluntary guidelines addressing the use and risks of artificial intelligence in election administration.
Unlike a mandatory regulatory system, the legislation specifically describes the guidelines as voluntary.
Artificial intelligence-generated content: Senate Bill 4915
Senate Bill 4915, the Artificial Intelligence Labeling Act, was introduced on June 24, 2026, by Senator Brian Schatz of Hawaii, with Senator John Curtis of Utah and Senator Mark Warner of Virginia as cosponsors.
The legislation would require disclosures for covered artificial intelligence-generated content and contains requirements directed at providers of generative artificial intelligence systems that produce covered content.
It was referred to the Senate Committee on Commerce, Science, and Transportation.
For businesses involved in generative artificial intelligence, advertising, digital media or content creation, disclosure requirements could become an operational issue if legislation of this type advances.
Competition and artificial intelligence agents: Senate Bill 5051
Senate Bill 5051, the Artificial Intelligence Access, Gatekeeper Exchange, and Nondiscriminatory Transfer Act, was introduced on July 21, 2026, by Senator Mark Warner of Virginia.
The bill's stated purpose is to promote competition and reduce consumer switching costs in the provision of online services.
It was referred to the Senate Committee on Commerce, Science, and Transportation.
The proposal represents another dimension of the artificial intelligence debate: competition and consumers' ability to move between online services.
Federal government use of artificial intelligence: Senate Bill 5057
Senate Bill 5057, the Safeguarding Against Fabricated Exploitation Through Artificial Intelligence Act, was introduced on July 21, 2026, by Senator Mark Warner of Virginia.
The legislation would prohibit the use of certain artificial intelligence models across the federal government.
It was referred to the Senate Committee on Homeland Security and Governmental Affairs.
This proposal focuses specifically on federal government use of artificial intelligence rather than establishing rules for the entire private sector.
A fragmented federal approach.
The House and Senate proposals demonstrate that lawmakers are approaching artificial intelligence from substantially different directions.
The House FRONTIER Act concentrates on federal oversight of advanced artificial intelligence developers.
The Senate proposals address individual sectors and problems:
- financial-sector experimentation;
- evaluation of advanced artificial intelligence risks;
- workforce forecasting and artificial intelligence-related employment changes;
- election administration;
- disclosure of artificial intelligence-generated content;
- competition and consumer switching costs; and
- the use of certain artificial intelligence models by the federal government.
These are not provisions of one comprehensive federal artificial intelligence law. They are separate bills, introduced by different lawmakers and referred to different committees.
None of the proposals discussed here should be described as current federal artificial intelligence law.
What the legislation means for businesses.
For business leaders, the most immediate lessons may come from operations rather than Washington.
The Supply Chain Dive analysis suggests that businesses need to establish a practical foundation before expanding artificial intelligence use.
That means identifying a decision where artificial intelligence can provide a measurable advantage, confirming that the necessary information is available and determining how employees will interact with the system.
It also means establishing a method for determining whether the investment is working.
That can involve measures such as forecast accuracy, inventory turnover, logistics costs or time saved, depending on the application.
The policy debate adds another layer.
Businesses developing advanced artificial intelligence systems will need to follow developments surrounding frontier-model oversight. Financial institutions should monitor proposals affecting artificial intelligence experimentation. Employers should watch legislation concerning artificial intelligence and workforce reporting. Companies producing or distributing artificial intelligence-generated content may have an interest in labeling legislation.
The result is an increasingly complex business environment in which artificial intelligence strategy and public policy are beginning to intersect.
The bigger business question.
The federal government has not yet settled on a single model for artificial intelligence regulation.
The House of Representatives is considering targeted oversight of advanced artificial intelligence developers. Senators are pursuing separate proposals dealing with specific industries, government functions, workers, consumers, competition and artificial intelligence-generated content.
Meanwhile, companies are already making decisions about whether artificial intelligence can improve their operations.
For businesses, that means waiting for Washington to resolve the broader policy debate is not necessarily an option.
Companies still have to decide where artificial intelligence belongs in their operations, how its performance will be measured and when an unsuccessful project should be discontinued.
The most immediate lesson from the supply-chain sector may therefore be one that does not depend on which artificial intelligence bills eventually become law:
Businesses need to establish the problem, the data, the performance standard and the expected return before they scale the technology.
The House of Representatives and the Senate are debating how the federal government should respond to increasingly capable artificial intelligence. Businesses, meanwhile, are already determining whether the technology deserves a place in their operations.
