House subcommittee hears how AI and data tools can speed recovery of trafficking victims

Chairwoman Mace convened the Subcommittee on Cybersecurity, Information Technology, and Government Innovation to examine how technology can help detect and prosecute human trafficking. Witnesses from Polaris, the National Center for Missing and Exploited Children (NCMEC), Marinus Analytics and Howard University described tools that have accelerated rescues and strengthened cases, while also flagging limits and harms if technology is deployed without guardrails.

Megan Lundstrom, chief executive officer of Polaris and a survivor of trafficking, told the panel that technology changes traffickers’ methods but does not change their motives. Lundstrom said anti‑trafficking systems must follow three principles: center survivor autonomy and informed consent, protect privacy through strong governance, and pair algorithms with human expertise. “Technology should never be used on survivors. It should be used with and for us,” Lundstrom said.

Melissa Snow, executive director of child trafficking programs at the National Center for Missing and Exploited Children, described how NCMEC uses donated tools such as Traffic Jam, Spotlight and Traffic Grama to match missing‑child reports and images with online advertisements and other open‑web indicators. Snow said NCMEC has responded to “over 280,000 reports” related to child exploitation and that, after the REPORT Act required platforms to report child‑trafficking indicators, NCMEC saw a dramatic increase in reports—she testified to a 952% rise in the first six months after the law took effect.

Cara Jones, co‑founder and CEO of Marinus Analytics, described Traffic Jam’s commercial ad index and its role in multi‑jurisdiction takedowns. She told lawmakers the platform has indexed more than 1.3 billion records and that, in a 2019 investigation, agents identified and recovered multiple victims and traced bookings to about 30,000 unique customers. Jones said automated screening has helped investigators link local cases to national networks and “move detection and recovery from months to days.”

Panelists told members that the value of these technologies depends on training data, human review and appropriate limits. Roy Austin, director of the Artificial Intelligence Initiative at Howard University, cautioned that systems trained on incomplete or biased data can misdirect investigations and exacerbate racial or gender disparities. He urged Congress to require transparency, auditing, limits on data retention and clear access rules.

Lawmakers pressed witnesses on forensic handling, data access and how NCMEC becomes involved after a law‑enforcement report. Snow described intake procedures and said NCMEC receives tens of thousands of reports annually; she also said many missing‑child records exist in law‑enforcement databases such as NCIC. Lundstrom said Polaris has developed hundreds of protocols to decide when to report information to law enforcement and emphasized that survivors’ consent and trauma‑informed practices should guide data collection and sharing.

The subcommittee did not vote on legislation at the hearing. Chairwoman Mace asked members to submit materials and written questions within five legislative days. Ranking Member Brown concluded by urging bipartisan work to pair technology with social supports and trained investigators so tools protect survivors without causing harm.