Privacy in Peril: AI and Government Surveillance

The U.S. military is seeking to enhance surveillance capabilities using advanced AI technologies, raising significant concerns about the monitoring of citizens’ movements, online activities, and personal connections. This emerging conflict between the Department of Defense and leading AI firms underscores the potential for AI-driven surveillance to be misused against the American public and highlights the pressing need for Congressional oversight.

Recently, the Pentagon and former President Donald Trump announced their decision to discontinue the use of Anthropic’s AI products. They criticized the company’s proposed safety measures—such as prohibitions on mass domestic surveillance and fully autonomous weapons—as insufficient. The Trump administration went further, declaring these stances rendered Anthropic a ” supply chain risk“, and barred any entity working with the military from engaging with Anthropic.

However, this issue transcends a typical contract dispute; it is fundamentally a struggle over the future of privacy in America, which will impact us all.

Central to this controversy is the government’s claim that it should have the right to utilize AI for any “lawful” purpose. The challenge lies in the law lagging significantly behind technological advancements. Current regulations do not adequately address a reality where smartphones act as tracking devices and our online behavior reveals as much as a diary. Data can be traded publicly, and AI enables seamless integration of purchased data into a vast compilation of domestic profiles.

This dilemma is exacerbated by the executive branch’s tendency to define “lawful” in secret. Without explicit rules from Congress, the Trump administration could validate a domestic spying initiative merely by asserting its legality. Given the government’s relentless appetite for data—and its history of sidestepping the Fourth Amendment’s protections against unreasonable searches—this is a truly alarming prospect.

Currently, the Department of Defense and various federal agencies already claim the authority to “lawfully” acquire Americans’ personal information—such as location and internet browsing history—without requiring a warrant. Despite bipartisan condemnation in Congress regarding these warrantless searches, efforts to curtail these practices have thus far been unsuccessful.

The situation is likely to deteriorate further with the advent of AI. Reports from the New York Times indicate that the Pentagon intends to leverage AI to collect and analyze large sets of unclassified, commercially-available data on citizens, including geolocation and web-browsing records. This revelation not only confirms that the government actively gathers vast amounts of private data but also reveals its ambition to use powerful AI tools to exploit this controversial information.

AI technologies could empower the government to derive insights about an individual at the click of a button, insights that previously required significant time and effort. These systems can consolidate information from various sources, detect patterns, and synthesize a comprehensive profile of a person’s activities and beliefs. For instance, if the government buys a dataset reflecting the movements of numerous cellphones, it may lack identifiable names, necessitating further analysis. AI can swiftly execute this analysis, integrating other data streams to create a fuller picture of an individual’s life at an unprecedented scale, which is particularly concerning given the Trump administration’s rush to access sensitive data like voting records, health information, and tax details.

While Anthropic navigates its challenges with the Pentagon, other federal agencies also purchase private data in bulk. The ACLU’s Freedom of Information Act inquiries have revealed that ICE has repeatedly acquired cellphone location data and license plate information to target immigrant communities. Moreover, federal agencies have begun collecting license plate data and facial recognition data from protesters and others documenting public activities. In light of this context, the extent of power that these agencies are amassing through AI is deeply concerning.

AI facilitates instantaneous surveillance, eliminating human barriers and amplifying the risks of digital spying by increasing speed, affordability, and thoroughness. If Congress does not take action, the application of AI to “lawfully” acquired data could lead to the creation of an extensive government database filled with sensitive information about every citizen. The ramifications of an AI-informed domestic database are dire, potentially resulting in widespread invasions of privacy, severe restrictions on free speech and assembly, and the increased targeting of marginalized or unpopular groups.

The overall impact on society is corrosive. Experience shows that government surveillance can lead to discriminatory practices and profiling, resulting in unjust investigations and prosecutions that leave lingering fears for years.

On Monday, OpenAI—after announcing an agreement with the Pentagon following Anthropic’s setback—stated it was modifying its deal to include protections for the civil liberties of U.S. citizens and permanent residents. Although this move is commendable, the new provisions are filled with loopholes. It brings to light a larger issue: individual rights shouldn’t hinge on the preferences of a single CEO. The questions surrounding government acquisition of Americans’ private information and the use of AI to analyze it are highly significant. Their answers shouldn’t rest upon individual contracts subject to change or the private interests of tech executives. The American public deserves a meaningful and enduring legislative framework to safeguard their privacy. It is imperative that Congress takes action.

Congress can begin this vital effort by passing the bipartisan Fourth Amendment Is Not For Sale Act, a common-sense reform initiative prohibiting the government from purchasing data that it would need a warrant to access otherwise. Additionally, there is an urgent need to implement fundamental safeguards on the government’s use of innovative AI technologies, ensuring protections against warrantless privacy invasions and intrusions on freedoms of speech and assembly, both online and offline.