Source: Amnesty International –
The US authorities are using automated artificial intelligence (AI)-powered surveillance tools to deliberately target non-US citizens and pose risks to those who speak out for Palestinian rights, said Amnesty International, amid the country’s ongoing unlawful clampdown on migrants, refugees and asylum seekers.
Amnesty International reviewed documentation from the Department of Homeland Security’s (DHS) public records, and previously disclosed procurement and privacy assessment documents, showing that AI tools Babel X, provided by Babel Street, and Palantir’s Immigration OS, have automated capabilities that enable constant mass monitoring, surveillance, and assessments of people, often for the purpose of targeting non-US citizens. The research also exposes how these tools are being used by the US government to track migrants, refugees and asylum seekers, and ultimately carry a high risk of being used as part of the “Catch and Revoke” initiative.
“It is deeply concerning that the US government is deploying invasive AI-powered technologies within a context of a mass deportation agenda and crackdown on pro-Palestine expression, leading to a host of human rights violations. These technologies enable authorities to swiftly track and target international students and other marginalized migrant groups at an unprecedented scale and scope. This has led to a pattern of unlawful detentions and mass deportations, creating a climate of fear and exacerbating the ‘chilling effect’ for migrant communities and for international students across schools and campuses,” said Erika Guevara-Rosas, Senior Director for Research, Advocacy, Policy and Campaigns at Amnesty International.
The State Department’s repressive AI-driven “Catch and Revoke” initiative combines social media monitoring, visa status tracking, and automated threat assessments of foreign individuals on visas, including international students. The AI tools supplied by Babel Street and Palantir, play a key role in US authorities’ ability to identify people, as well as their behaviour and movements, more quickly and on a large scale to determine if their visas should be revoked.
Babel X conducts sentiment analysis by assigning a sentiment and likely intent to posts shared by individuals based on online behaviour, while both Babel X and ImmigrationOS can conduct pattern recognition and automate data analysis using algorithms. The tools also aggregate data from various public and private sources, including multiple government databases.
The coercive ‘Catch and Revoke’ initiative, facilitated by AI technologies risks supercharging arbitrary and unlawful visa revocations, detentions, deportations and violations of a slew of human rights. These include the rights to privacy, freedom of expression and access to information, freedoms of movement equality and non-discrimination, and the right to liberty and protest.
Erika Guevara-Rosas, Senior Director for Research, Advocacy, Policy and Campaigns
Babel X, for instance, is allegedly used for AI-powered scanning of social media platforms for flagging “terrorism” related content. This information can be used by US authorities to make a decision about revoking an individual’s visa. If the State Department determines a visa to be revoked, ICE agents are dispatched from local field offices to arrest and deport the individual. Probabilistic technologies that are used to draw inferences about individuals’ intent have massive margins for error, and can often be discriminatory and biased, and could lead to falsely framing pro-Palestine content as antisemitic.
The human rights risks of such systems are well-known and longstanding. During the first Trump administration, 55 human rights groups (including Amnesty International), as well as 54 technology experts, wrote to the authorities to oppose the introduction of a similar automated “extreme vetting” system, on the grounds that it would be ineffective and invariably lead to violations of the right to non-discrimination, expression and other human rights of migrants in the USA.
“The coercive ‘Catch and Revoke’ initiative, facilitated by AI technologies risks supercharging arbitrary and unlawful visa revocations, detentions, deportations and violations of a slew of human rights. These include the rights to privacy, freedom of expression and access to information, freedoms of movement equality and non-discrimination, and the right to liberty and protest,” said Erika Guevara-Rosas.
Amnesty International analysed procurement documents, contracts, and Freedom of Information Act (FOIA) responses to previous requests by other organizations, related to all publicly known AI projects currently deployed by the US Department of Homeland Security (DHS), Custom and Borders Protection (CBP), and Immigration and Customs Enforcement (ICE). According to the CBP website, ICE and CBP currently host at least 80 AI projects.