Artificial Intelligence in Immigration

ARTIFICIAL INTELLIGENCE IN IMMIGRATION

By: Laurie Snider, Fragomen, Del Rey, Bernsen & Loewy
Overview
Artificial intelligence (AI) has undoubtedly become the fastest growing and most prominent disruptive factor in many environments in the last two years. AI advancements present opportunities for several benefits in the immigration space, including streamlining the recruitment of foreign workers, reducing application processing times, enhancing security measures, and the ability to predict migration patterns. However, inherent challenges and risks of recent technology, including potential biases, lack of transparency, and data privacy issues, must also be considered.
The most common uses of AI in immigration systems are illustrated below:

This data reflects responses from a survey of Fragomen’s top 50 countries by volume on the use of artificial intelligence by government bodies within immigration systems. Figures represent the total count of countries that answered “yes” or “sometimes” for each category. Information was gathered in October 2024 and may change over time.

Benefits of AI in the immigration space
■ Streamlining the recruitment of foreign workers. AI can help with talent recruiting, creating job requirements, sourcing candidates and filtering resumes. As a result, the human time and effort devoted to not only hiring in general, but also to labor market testing by an employer or immigration practitioner, depending on the jurisdiction, could be greatly reduced.
■ Better use of government workers’ time in immigration departments. AI is more
efficient and quicker at performing repetitive tasks, such as reviewing large numbers of documents and easily identifying when a required document is missing from an application package. Additionally, it has widely been accepted that generative AI technology is likely to have the most impact in automating some tasks while leaving time for other valuable duties, as opposed to fully automating jobs and eliminating positions. Therefore, AI programs can help reduce easier tasks during the immigration application administration process while allowing immigration department staff to redirect their time to complex case analysis or assistance.
■ Positive impact on the labor market. While initial concerns focused on AI eliminating jobs, it is increasingly evident that AI will likely create a net positive of jobs by increasing worker productivity. According to Goldman Sachs Research, AI could boost global GDP by 7% and labor productivity by 1.5% over the next decade. This growing demand for AI-skilled workers is already apparent, with individuals proficient in AI earning already earning 25% higher salaries in markets like the United States. This surge in demand for AI talent creates opportunities for immigration policies to attract skilled workers.
■ Enhancing security measures. AI can also be used to enhance security, monitoring and
detecting potential breaches in real time through monitoring and/or surveillance software, such as for visa-overstayers or work permit constraint violators.
■ Assisting migration management with predictive technology. Perhaps most significantly,
governments that work in collaboration with private and public stakeholders can pool information to better plan for migration events, create managed migration policies, and predict migration patterns or activities.

Challenges and risks of AI in immigration
While AI holds the promise of streamlining immigration processes and enhancing efficiency, it is crucial to address the accompanying risks and ethical dilemmas. A key to mitigate these challenges lies in preserving the human touch within immigration processes. This human-centric approach can be preserved and implemented at various levels, from government agencies ensuring that human oversight remains integral to AI-powered processes, to companies engaging professional immigration services that leverage the human element.
Key challenges and risks to immigration departments and border management departments adopting AI include the below:
■ Biases in AI systems. The United Nations Committee on the Elimination of Racial
Discrimination has warned that AI and facial recognition technologies could reinforce racial bias and xenophobia, leading to human rights violations. The UN Committee also highlighted the risk of AI systems perpetuating existing biases if trained on inherently biased data. Potential biases can also extend to geographical disparities. For instance, individuals residing in technologically advanced countries or regions may have an undue advantage, such as facilitated access to appointment scheduling that will not be available to others.
■ Transparency and accountability. Technical challenges, such as faulty data used for AI training or issues in system upkeep, can affect the precision and dependability of AI. A notable concern is the absence of human decision-makers. This lack of the human touch (and lack of sourcing information in AI- produced text, known as the ‘black box’) can complicate the process of challenging or appealing decisions. Appellants may not be able to identify or recover the information on which decisions were based, leaving them without the ability to challenge interpretations or sources of information. Moreover, it introduces accountability issues, as pinpointing responsibility becomes difficult when mistakes occur.
■ Data privacy concerns. AI used in border management functions, refugee integration
projects, crisis detection, speech recognition in asylum procedures, and identity fraud detection has led to debates about the protection of personal data and the potential misuse of this data. Though some government agencies are attempting to protect against their vulnerabilities because of the rise of AI in their systems (as well as adversarial attacks against their systems), this is an area that requires further development. This risk in particular may limit AI’s use in immigration as governments and employers must remain compliant with data privacy laws or may be unwilling train systems on particularly sensitive or proprietary topics.
■ National security issues. If governments choose to utilize AI to perform routine
immigration functions, they will have to be careful which AI platforms they choose to host sensitive and proprietary governmental information, due to national security implications.

■ Translation issues. As the frequency of AI being used in the form of neural machine
translations increases, there can be significant consequences in the immigration landscape, where immigration decisions can often turn on a few words or sentences. This became an issue for Afghan asylum seekers in the United States where errors in machine translations of Pashto and Dari were causing written statements by asylum seekers to not match interview statements, resulting in the denial of applications.
What’s next?
■ Potentially more restrictive immigration policies. As AI use spreads, governments are likely to implement restrictive trade policies concerning AI, including similarly restrictive immigration clauses. This will come as a result of AI being used to share sensitive data, including national security information; for example, governments are using AI to predict failures in weapons platforms. As a result, friendshoring is likely to increase.
o However, this approach to keeping sensitive AI technology with “like-minded” countries may be more complicated than initially anticipated, since certain countries could have vested interests in both sides of global rivalries.
o Alternatively, companies could be banned from operating in certain countries perceived as being against their operating country’s interests, or strategic deals could be made that include exclusivity provisions.

■ AI developments could result in increased friendshoring and nearshoring practices, creating hotspots for needed talent. As AI develops, governments are likely to increase friendshoring (when a manufacturer or service provider moves all or part of their business
to another geographical location, usually one with which it has political alliances) and nearshoring practices (when a manufacturer or service provider moves all or part of their business to another geographical location, usually one relatively close to the company’s headquarters), for various reasons. First, as noted above, AI is used in the production of items such as semiconductors, which are used in almost every developed country’s defense systems. The production of semiconductor chips used in military products in adversarial countries could potentially have devastating national security implications with top secret
information being leaked. As such, governments may seek to limit which businesses or governments are able to access these products or information.
■ Re-composition of workforce. With the emergence of AI, many lower-skilled jobs are likely to disappear. If AI replaces these lower-skilled workers, the demand for foreign nationals to take these jobs will decrease. On the other hand, various forms of automation throughout history have actually created new types of jobs; AI could potentially create long-term economic growth in countries that are early adopters of the technology. Thus, although AI may decrease the number of lower-skilled migrants needed, it is likely to increase the demand for those skilled in AI-related fields.
■ Skills gaps create need for reskilling and new immigration policies. Companies will need to become innovative in their approach to reskilling as many of their employees’ job functions will be replaced by AI. Further, in order to find the right candidates for these AI- related jobs, organizations are more attentive to the skills that candidates possess, rather than their degrees. By decreasing this major barrier to employment, more candidates around the world, particularly from developing nations where there are fewer opportunities for formal education, will be able to access the labor market for these new AI jobs.
Governments, too, are beginning to recognize this need to reevaluate educational requirements in order to fill labor shortages in areas such as AI.
■ Increased private sector involvement in policy decisions. As governments seek to regulate AI, the private sector’s expertise in this field will be critical to drafting policy that allows for innovation and productivity while avoiding pitfalls of the technology, such as data privacy breaches and national security issues.
Conclusion
The effect of AI on immigration policy, migration patterns, and immigration alliances are undeniable, though it remains to be seen just how much the immigration environment will be impacted by AI. This is a trend we are tracking in our Worldwide Immigration Trend Reports.
For more on this topic, see this Fragomen blog.

For additional questions on this topic, or other corporate immigration issues, please contact Laurie Snider at lsnider@fragomen.com.
The information contained herein is current as of January 2025.