Virginia legislators are considering imposing a complex compliance regime on nearly everyone who touches artificial intelligence for a wide range of business uses. House Bill 2094 would affect not only well-funded firms like OpenAI or Google DeepMind and small AI startups, but nearly every Virginia business. In some contexts, it would even affect individuals.
It is, in other words, a staggeringly broad law that would push Virginia in the direction of the heavily regulated European Union.

Dean Ball
No one should sugarcoat the fact that AI creates major challenges to society, even if you believe the benefits will greatly outweigh the risks. It may meaningfully enable large-scale cyberattacks and other catastrophic harms. It could upset the balance of power in our republic in ways that require intervention of some kind. And it could eventually result in rapid reductions in human employment in some industries, necessitating at least short-term relief for workers. However, HB 2094 addresses none of these challenges head-on.
People are also reading…
What the bill does target is “algorithmic discrimination,” or the notion that an autonomous system of some kind might discriminate against people based on age, religion, gender, race, or other civil rights-protected traits. It does so primarily with paperwork burdens and legal liability when an AI system is used to make a “consequential decision” that could have a “material effect” on a consumer’s access to things like insurance, financial services, employment and education, or the terms of such agreements.
These requirements apply to AI developers, as you might expect. They apply to “integrators” who combine AI with other software systems, and to cloud computing companies which distribute AI systems. Finally, they apply to “deployers,” which includes anyone who uses an AI system. And whether or not you are a fan of AI, its use will become much more prevalent and necessary to make a living moving forward.
Imagine being the owner of flower shop who needs to hire a new employee, or even a parent hiring a nanny. You post the job online and get a few dozen applications. Lacking time, you simply ask ChatGPT to pick out the applicants who have previous, relevant experience.
In most of America, that’s a normal and acceptable use of AI. However, in Virginia, that would make you a “deployer” of a “high-risk artificial intelligence system” used as a “substantial factor” in making a “consequential decision.”
As a result, you’d need a “risk management plan” and “algorithmic impact assessment.” The latter would need to be updated each time there is a “substantial modification” made by OpenAI to ChatGPT, which occurs roughly once per month. If you did not write these documents (or more likely, pay a consultant to write one for you), you’d face a fine of up to $1,000.
And if your system is found to have had a disparate impact on any protected class, regardless of whether you — someone who merely used a publicly available platform — intended to discriminate, you’d be subject to lawsuits.
This is merely one example; just as ChatGPT and other frontier language models have countless productive uses, the range of ways this law could be applied boggle the mind. Given how large AI’s presence in our lives is likely to become, the burdens this law imposes will only grow with time.
This is because AI is a general-purpose technology, much like electricity, the integrated circuit, or computers themselves. The difference is that we do not make businesses or individuals fill out paperwork to use electricity.
What’s more, the kind of discrimination HB 2094 envisions is already illegal under both Virginia and federal law. If a business discriminates against a particular group in its hiring decisions, that’s a crime, regardless of whether pen and paper, a computer, or an AI system was used. What HB 2094 really adds to the equation, then, is a mountain of compliance paperwork and a boon for consultants and lawyers.
Laws governing the use of general-purpose technology should be intuitive for the average user, whether it’s an individual or corporation. Members of the state legislature should be thinking about whether this bill is intuitive enough, and whether imposing so many complicated mandates on their constituents will weaken Virginia’s economic competitiveness. They should also consider whether this move would push America’s digital economy further toward the same kind of regulatory morass in the physical world that has made it so difficult to build new infrastructure.
The recent election showed us that Americans desire a simpler regulatory regime, not a more complicated one. Our state lawmakers should take this to heart, though it is no small task. Simplicity may be far harder to achieve than complexity, but it’s worth the effort.
States Whose Businesses Are Embracing Artificial Intelligence
States Whose Businesses Are Embracing Artificial Intelligence

Photo Credit: Nuttapong Punna / Shutterstock
As artificial intelligence technologies have become more prevalent in the last few years, businesses everywhere are trying to understand what AI means for them.
Major tech corporations like Google, Microsoft, and OpenAI (maker of ChatGPT) have gotten headlines for driving the development and adoption of artificial intelligence, but AI has the potential to impact companies of all sizes and sectors, including small and local businesses. From automating routine tasks to improving efficiency and personalizing customer service and marketing, businesses are already identifying a variety of uses and applications for this emergent technology.
However, nationwide data shows that AI is still early in the adoption curve, and businesses in the U.S. still have a long way to go in integrating and maximizing the benefits of artificial intelligence.
Trends in AI Adoption in the Workplace

A small but growing share of U.S. businesses are using AI
Source: Tivly analysis of the U.S. Census Bureau’s Business Trends and Outlook Survey data
Despite all the buzz, AI adoption in business remains fairly low. According to data from the U.S. Census Bureau, only around 5.5% of businesses report that they are currently using AI in the production of goods and services. However, AI adoption has been steadily increasing and this trend is expected to continue. Just 3.7% of businesses were using AI 12 months ago, but 8.1% of firms report that they plan to use AI in the next six months.
Business Use of AI by Firm Size

Small businesses are more likely to use AI to replace employee tasks
Source: Tivly analysis of the U.S. Census Bureau’s Business Trends and Outlook Survey data
So far, larger businesses tend to be most likely to embrace AI technology. Among the largest firms—those with more than 250 employees—8.3% report using AI to produce their goods and services. Interestingly, however, the very smallest businesses also appear open to integrating AI into their work.
Small businesses with one to four employees have the third-highest rate of AI adoption at 6.1%. These smaller companies are also the most likely to use AI to replace employee tasks. A total of 28.5% of AI-adopting firms in this size group report doing so—the highest across all business sizes. For small business owners, AI may be an especially useful resource for completing routine but often time-consuming tasks like data entry or simple communications, allowing them to improve productivity and efficiency despite their low headcounts.
Artificial Intelligence Adoption Among U.S. Businesses

Source: Tivly analysis of the U.S. Census Bureau’s Business Trends and Outlook Survey data
In addition to business size, rates of AI adoption also vary by geography and industry. Firms in the South and Midwest tend to have the lowest rates of AI adoption, led by Mississippi at just 1.7% of businesses. Western states have higher rates of adoption than most of the country, with Colorado topping the U.S. at 7.4% of businesses, while select Eastern states like Florida and Delaware also rank highly.
Some of the geographic divide may be driven by industrial factors. Fields like information (20.5%); professional, scientific, and technical (14.9%); and educational services (10.3%) are most likely to use AI, and those industries are also most common in the regions of the country where AI use in business is high. The fields that use AI the least include construction, transportation and warehousing, and accommodation and food services, all of which have less than 2% of firms using AI.
The analysis was conducted by Tivly using data from the U.S. Census Bureau’s Business Trends and Outlook Survey. For the complete analysis, see States Whose Businesses Are Embracing Artificial Intelligence on Tivly.
Methodology

Photo Credit: Nuttapong Punna / Shutterstock
To find the states whose businesses are embracing Artificial Intelligence (AI), researchers from Tivly—a commercial insurance marketplace and services provider—used the latest data from the U.S. Census Bureau’s 2024 Business Trends and Outlook Survey. The researchers ranked states by the share of employer businesses that used AI in the production of goods or services in the previous two weeks. Some examples of AI are machine learning, natural language processing, virtual agents, and voice recognition. Researchers also calculated the share of businesses using AI to perform tasks previously done by employees (among businesses that use AI), the share of businesses planning to use AI in the next six months, the share of businesses planning to use AI to perform tasks done by employees (among businesses planning to use AI in the next six months), and the most common type of AI used among businesses. To reduce variability and to supplement the primary data fields, the latest data available from the Business Trends and Outlook Survey’s AI Supplement dataset was used for certain national and state calculations.
For complete results, see States Whose Businesses Are Embracing Artificial Intelligence on Tivly.
Dean W. Ball is a research fellow with the Mercatus Center at George Mason University in Arlington. Contact Ball at dball@mercatus.gmu.edu.