Portman introduces two bills on facial recognition, AI in government
Sen. Rob Portman (R-Ohio) introduced two new bills on Wednesday establishing civil rights guardrails for the use of facial recognition and artificial intelligence technology. Anna Moneymaker / GETTY IMAGES
By
Natalie Alms
The fate of the bills in a lame duck session of Congress is uncertain, but a Portman aide says the outgoing senator wants to be active on the subjects even after retiring.
Outgoing Sen. Rob Portman (R-Ohio) introduced two new bills on Wednesday meant to guard against civil liberty and civil rights violations in the government’s use of facial recognition technology and artificial intelligence.
The Facial Accountability, Clarity and Efficiency In Technology Act, or FACE IT Act, wouldn’t preclude the government’s continuing use of the tech, but it would set up guardrails, Portman said in a statement.
The bill would have the National Institute of Standards and Technology set out minimum accuracy requirements for government use cases of the tech and bar agencies from procuring facial recognition technology that didn’t meet those standards.
The proposal would also add restrictions on government access to facial recognition databases, require a human to be involved in decision-making by the tech and give an opt-out and alternative option for citizens in instances like identity verification.
“Facial recognition technology can be used to help protect our communities, but I am concerned about the potential for abuse,” Portman said. “Given the civil liberty implications of the federal government’s use of facial recognition technology, we must pass legislation to set rules for the use of this technology.”
Although the use of facial recognition technology has become increasingly common in the government for use cases in national security, digital identity and law enforcement, advocates have raised alarm bells about the potential for surveillance and tracking and the implications for equity and civil rights, particularly around the use of face recognition in the criminal justice system and the existence of demographic differentials in some algorithms.
In 2021, over 60 civil rights and community organizations penned a letter asking Congress to prohibit the government's use of facial recognition technology.
Although Congress has not passed any comprehensive legislation on the use of facial recognition in government, there have been several other proposals around the tech, some curtailing its use in state and federal government and others focused on the use in law enforcement.
The privacy bill considered by lawmakers in this Congress also had protections around the collection and sharing of biometric data.
Portman, co-chair of the Senate Artificial Intelligence Caucus and outgoing ranking member of the Homeland Security and Governmental Affairs Committee, also introduced a bill focused on artificial intelligence.
The proposal, called the Stopping Unlawful Negative Machine Impacts through National Evaluation Act, would “clarify that existing civil rights laws apply to decisions made by AI systems just as if those decisions were made by humans” by spelling out that civil rights laws apply to AI-enabled decisions, Portman’s statement about the proposal said.
The bill would also give NIST the power to set up so-called tech evaluations with stakeholders to find ways to reduce discrimination in algorithms as an alternative to AI regulations, Portman says.
“Rather than jump all the way to heavy-handed regulation of this new technology, Congress should make clear that we already have numerous laws on the books to guard against discrimination on the basis of protected characteristics,” Portman said in a statement about the bill.
A common complaint about artificial intelligence is the potential for systems to advance discrimination at scale if algorithms have biases baked into them, and issues around accountability for the systems when decision-making and outputs are difficult to understand.
Ben Winters, counsel at the nonprofit Electronic Privacy Information Center, told FCW via email that “specifically articulating that civil rights protections can’t be skirted by the use of AI is helpful and sends a message to people considering its use to make sure to do so responsibly,” although he added that he would “love to see this go further” with provisions such as explicitly allowing assessment findings to be referred to regulators.
The White House’s “AI Bill of Rights,” released this fall, included sections on algorithmic discrimination and notice and explanation about decisions made by automated systems.
The fate of Portman's new proposals is murky, given that the current Congress ends Jan. 3, 2023.
A Portman aide told FCW that the senator “wants to help drive a productive conversation around the responsible use of AI which is a debate that will continue long after he retires,” when asked about the timing of the bills’ introduction.