What happens when an immigration officer points a smartphone or another smart device at someone’s face and captures an image?
In this video, immigration attorney James Martin describes a process that begins with facial identification and continues with the analysis of information drawn from multiple databases. He discusses how artificial intelligence, facial recognition and tools associated with Palantir could affect immigration enforcement—and what happens when the information used is wrong.
His discussion addresses three central questions:
- How are immigration authorities using identification tools and data to learn about individuals?
- What are the shortcomings and risks of these technologies?
- What practical steps should people consider when dealing with an immigration matter?
Martin, who handles immigration cases from across the United States, also shares an example from his practice that illustrates why an attorney may need to intervene before a client answers questions based on incomplete or misleading information.
From a Facial Image to an Immigration Investigation
Martin opens with a scenario in which an officer captures an image of a person’s face using a smartphone or similar device. The image is then used to identify the individual through facial recognition.
In describing this process, he states that the technology relies on facial features that can be traced back to birth and that remain unchanged from infancy to old age. Once identification takes place, he explains, the agency can bring information into an artificial intelligence program to investigate the person’s background.
The categories of information he discusses include:
- Immigration or legal status.
- Travel history.
- Previous activities.
- Tax filings.
- Possible fraud.
- Prior encounters at the border.
Martin characterizes the potential reach of this process as allowing an agency to find out “virtually everything” about a person. He describes the resulting environment as a “brave new world.”
Does Artificial Intelligence Make the Final Immigration Decision?
Martin distinguishes between the use of artificial intelligence during an investigation and the final decision in an immigration case.
He explains that AI does not make the final decision. However, he argues that it can have a substantial impact on how a case is processed and what the government believes it can do.
In particular, he identifies three roles or characteristics:
- Identifying individuals.
- Obtaining information from databases that may be difficult to access through traditional methods.
- Producing errors.
He emphasizes that an AI system is only as good as its programming and training. The technology can make mistakes, and its output should not be treated as infallible.
Martin also notes that AI is available beyond government agencies. Nevertheless, he describes the government as fully committed to using it in enforcement activities.
Palantir, ImmigrationOS and Information Across Databases
Martin discusses a government contract involving Palantir and describes ICE as having disclosed that it is using, or will use, Palantir technology to relate information across multiple platforms almost instantaneously.
He identifies ImmigrationOS, or the Immigration Operating System, as a closely watched development. In his account, the system allows Immigration and Customs Enforcement officials to bring together information from different sources.
The Barriers Between Government Agencies
To explain why he considers this significant, Martin contrasts the current situation with what he recalls from earlier practice.
He describes an institutional barrier between government agencies, using the expression “Chinese wall.” His recollection is that, in the past, immigration officials seeking certain tax information would have needed to obtain it from the client or secure the client’s authorization.
Martin presents those barriers as privacy protections. In his view, keeping information and power distributed between agencies helped protect individuals from excessive government intrusion.
He argues that the ability to connect information across agencies changes that balance and places substantially more information within the reach of immigration enforcement.
Location Tracking, Cameras and Facial Recognition
Martin also describes capabilities involving location information and monitoring.
He discusses the possibility of identifying people in crowds during demonstrations, accessing information about border crossings and checkpoints in real time, and using geolocation to determine someone’s whereabouts.
His argument is that the proliferation of cameras could make it unnecessary for a person to carry a smartphone in order to be identified or located. He further describes the potential to identify people’s associations.
These are among the developments that lead him to invoke the idea of “Big Brother” watching.
Privacy, Government Power and the Risk of Error
For Martin, the issue extends beyond the efficiency of an enforcement investigation.
He raises concerns about an agency’s access to information concerning where people go, whom they associate with, their tax records and their private lives. In his view, concentrating that much information in government hands creates serious risks.
He expresses particular concern about the administration discussed in the video and its record of protecting individual rights. However, he also makes clear that his criticism is not limited to one administration or political party.
His broader argument is that the pursuit of power is a human issue. Information creates power, and the accumulation of information can create opportunities for abuse.
Although he describes the technology as generally accurate, he repeatedly emphasizes that it remains fallible. The consequences of that fallibility can be substantial.
The Main Risks James Martin Identifies
Martin introduces his discussion as involving approximately four basic problems. Across the video, he develops concerns about identification errors, inaccurate underlying information, a lack of transparency and the potential misuse of data.
1. Identification Based on Probability Can Be Wrong
Martin explains that identification systems work by evaluating data points, including features of a person’s face, and producing a result based on probability.
They are not completely accurate.
For him, the practical issue is what happens when a person is detained because of faulty information. Even a short detention matters, and the consequences become more serious when the error affects someone for days or weeks.
He views this as both a problem of technological reliability and an intrusion on individual rights.
2. Incorrect Information Can Remain in a Database for Years
A second concern is the quality of the underlying data.
Martin points out that people entered much of the information in existing databases before the widespread use of current AI tools. Human mistakes may therefore remain in those records.
He also notes that artificial intelligence itself can produce bad information. He refers particularly to publicly available large language models that are not specialized for a narrowly defined task with specific safeguards and controls.
His concern is that incorrect information entered into a database twenty years earlier may be incorporated into a new analysis. He questions the system’s ability to distinguish reliable information from unreliable information before using it to generate an output.
The consequences, he argues, can be severe for the person whose record contains the mistake.
3. The Sources and Safeguards Are Not Transparent
Martin questions where the government obtains its information, how it acquires that information and which databases it accesses.
He says he has seen information emerge in court that, in his experience, would not have appeared there two or three years earlier.
He describes some investigative practices as a search for information that can be used against an individual. His concern is that an agency could begin with an intended outcome—such as finding someone ineligible for a benefit—and then work backward through the available data.
He also criticizes what he sees as an absence of meaningful disclosure about the information being used and the safeguards governing its use.
4. Extensive Data Can Be Used to Control or Harm People
Martin considers the risks of misuse under a government acting with harmful intentions.
He asks what could happen during severe civil unrest, a major terrorist threat or circumstances approaching civil war. In those scenarios, he raises the possibility of information being used for blackmail, control, intimidation or the erosion of rights.
These are scenarios he presents to illustrate his concern about the concentration of power, rather than incidents involving the client whose case he later describes.
Enthusiasm for AI Should Not Replace Judgment
Martin compares the excitement surrounding artificial intelligence with the early experience of using the internet.
He recalls the feeling of sending an email overseas in seconds. He also compares technological enthusiasm to a young person discovering the power of driving a car for the first time.
His concern is that excitement can outpace understanding.
He describes people presenting themselves as experts without the background or education needed to evaluate the information they receive. In his view, the ability to recognize that something does not look or sound right remains essential.
Combining uncritical enthusiasm for technology with a desire for power, he argues, creates a dangerous situation.
Immigration Cases Still Involve People, Procedures and the Rule of Law
Despite these concerns, Martin emphasizes that technology does not replace the people and procedures involved in an immigration case.
He explains that individuals still have an opportunity to have their cases heard before the immigration authorities. He also stresses that government agencies are not all-knowing or all-seeing.
For that reason, he encourages people to work with legal professionals who handle these issues regularly.
His point is that practitioners can observe the information agencies are examining, identify potential vulnerabilities and review the broader history of a case.
He describes the developments he is seeing in practice as unusual, alarming and difficult to follow in real time.
Practical Considerations for Your Immigration Case
Work With an Experienced Immigration Professional
Martin’s first recommendation is to consult a professional when seeking an immigration benefit.
In his view, the process is no longer simply a matter of completing forms. Nor should someone assume that information generated by AI will necessarily be correct.
He recommends working with people who understand immigration rules and have practical experience in the field.
Martin acknowledges that keeping up is challenging even for professionals. He says attorneys are also adopting tools while learning from their interactions with agencies, the information officials examine and the issues that influence how a case progresses.
That experience, he explains, can help a client understand what may happen if a particular course of action is pursued.
Be Mindful of Your Social Media Presence
Martin advises viewers to consider the information they make available through social media.
His concern is that posting extensively can provide authorities with additional information.
He also acknowledges the limits of this advice: some information may already be available, and changing posting habits may come after it has been collected.
Keep Copies of Every Immigration Filing
James Martin describes clients arriving at his office without a clear understanding of what someone else has filed on their behalf.
He emphasizes the importance of retaining copies of all immigration submissions and supporting documents.
We also encourage clients to insist that the professional representing them provide copies and review the forms with them so they understand what is being said in their name.
He highlights this point particularly for people whose first language is not English.
Do Not Assume an Allegation Is Accurate
Martin returns to the possibility of incorrect records and mistaken identity.
He describes situations in which a client has denied an allegation, insisted that they were not present or had not committed the act in question, and ultimately obtained approval of the case.
His message is that a person who is certain an allegation is wrong should not simply accept it. He also emphasizes the importance of having professional legal support when responding.
Answering Questions, Remaining Silent and Avoiding False Information
Toward the end of the video, Martin discusses how to respond when government questioning could create legal difficulties.
He uses the warning that “the government knows everything” to emphasize caution, while elsewhere in the video explaining that agencies are not all-knowing and their information can be wrong.
He discusses the right to remain silent and states that the government bears the burden of proving a case against an individual. At the same time, he repeatedly warns viewers not to provide false documents, false testimony or false information.
He stresses that inaccurate statements can create problems many years later.
Martin also discusses the distinction between answering truthfully and volunteering additional information. Where questioning could expose a person to adverse consequences, he says decisions about speaking, remaining silent or disclosing information require the assistance of experienced legal counsel who understands the implications.
His client example illustrates why he considers that judgment important.
A Client’s Erroneous Voter Registration: Why Intervention Mattered
Martin recounts representing a twenty-year-old woman who was seeking a green card through her stepfather.
During the case, immigration officials began asking questions about whether she had registered to vote. Martin describes the issue as potentially carrying very serious consequences.
He stopped the questioning and told the officer that his client would not continue speaking.
According to his account, the officer responded that the woman was an adult and could decide for herself. Martin then told the client, her mother and her stepfather to trust his judgment.
He warned that proceeding without all the relevant information could result in deportation and the possibility that she would never be allowed to return to the United States.
The officer indicated that the application might have to be denied. Martin nevertheless proceeded to investigate the underlying facts.
What the Records Showed
Martin says the woman had been registered to vote, but that she had marked on the application that she was not a U.S. citizen.
He obtained a letter from the county confirming that the registration had occurred in error and that she had not misrepresented her citizenship.
He explains that the application ultimately worked out.
For Martin, the lesson is that allowing questioning to continue without checking the records could have led the client to admit to something she had not actually done.
The existence of a record did not, by itself, establish that she had made the misrepresentation the questioning appeared to suggest.
Discuss Your Immigration Questions With James Martin
Martin closes by reminding viewers that they have rights and emphasizing the value of experienced counsel who understands the issues involved.
His central concern throughout the video is the combination of powerful investigative tools and information that may be incomplete, incorrect or misunderstood.
James Martin handles immigration cases from across the United States. If you have questions about your circumstances or would like to discuss his services, use the contact link accompanying the video.