Minister Fraser explains AI’s role in IRCC application processing. At a recent press conference in Vancouver, Immigration Minister Sean Fraser told reporters, “We have increased the use of technology and sophisticated analytics and expedited our processes over the past few months”.
During Fraser’s speech, he mentioned applications for temporary residences under the family class sponsorship program while applicants wait for permanent residence. Technology and advanced analytics have impacted the processing of applications across all lines of business at IRCC.
Immigrants are expected to flood Canada in the upcoming years in the largest number ever. It is twice as many decisions as in 2021 that were made by the IRCC on applications submitted across all business lines in 2022.
Furthermore, in recent years, Immigration and Refugee Protection and Customs Enforcement have rapidly shifted towards digitization and modernizing Canada’s immigration system. It has included increasing the use of AI – (artificial intelligence) and advanced data analytics to speed up application processing. Applicants can now anticipate a 30-day processing time for spouse TRV applications, as 98% of applications were processed by Fraser.
IRCC asserts that digitizing services and systems will fulfill the requirements of clients and Canada by using new technologies. Technology also allows the agency to process applications more quickly. As well as allows officers to focus on more complex applications by utilizing technology.
IRCC applies artificial intelligence (intelligence obtained through computer technology) to a number of processing areas, including:
- The automated determination of eligibility for benefits.
- Identifying characteristics of applications and distributing them between officers.
- Assessing whether additional verification is required for applications.
- Distribution of workloads.
- As a result of the creation of “annotations”, they reduce the number of officer searches in our global case management system for each client.
- Answering client emails in a timely manner and providing publicly available information in response to client inquiries.
- Biometric assessment.
An application cannot be approved or rejected by any of the tools used by the IRCC. It is not sufficient for a computer decision to grant or deny a candidate’s visa application.
What are the areas of concern regarding AI’s usage in IRCC application processing?
It has been expressed by a number of people that the IRCC’s use of artificial intelligence (AI) may lead to prejudice and a general lack of clarity regarding how applications are assessed.
Government-wide policy statements, such as the Treasury Board statement on Automated Decision-Making, warn that algorithms based on historical data can exacerbate disparities in racial, social, and gender outcomes. It states that not all skin tones or genders respond to facial recognition algorithms in the same way.
When AI is used too heavily, it may lead to a lack of clarity in decision-making. As per the directive, federal agencies must describe their administrative decision-making processes. In addition, those who are denied benefits or services have a right to clear and reasonable explanations from the government rather than just the assertion that a machine decided for them.
The IRCC has also faced criticism for its use of the “Chinook” tool, which it describes as a Microsoft Excel tool. Its purpose is to evaluate applications for temporary residence visas, work permits, and student permits. Chinook’s decision-making algorithms do not include any AI, advanced analytics, or algorithms built into them, according to the department.
What is the process by which IRCC develops algorithms?
According to IRCC, it complies with a directive in the Canadian Charter of Human Rights and Freedoms that guarantees equal rights and prohibits discrimination. As part of how decisions are made and how justification is given to those who are affected. It claims to adhere to standards of openness, responsibility, legality, and procedural fairness.
Every algorithm that is proposed must undergo an Algorithmic Impact Assessment (AIA). Currently, AIAs are used to assess risk in accordance with the Treasury Board’s Directive on Automated Decision-Making. The instrument assesses risk factors, mitigation, and the effect of suggested algorithms.
Other requirements include analyzing algorithms’ effects, ensuring data and algorithms are of high quality, and disclosing algorithms’ use in advance.
IRCC informs us that the rules used to support (but not finalize) the decision-making process are reviewed regularly by senior decision-makers, experienced officers, and legal, policy, and data science. And privacy experts to ensure they meet the Immigration and Refugee Protection Act’s requirements.
Moreover, regular monitoring and quality assurance techniques are also implemented to ensure that these models continue to work as intended. As well as that any unanticipated adverse effects can be identified and minimized early.
It remains an upbeat vision for how the IRCC will use modern technologies and sophisticated analytics to carry out its mission. A number of aspects of an applicant’s application can be examined by the IRCC using technology, according to him. Moreover, allowing the department to assess whether an applicant is likely to qualify for permanent residence. Thereby placing them in a category where IRCC officers are more likely to approve them. The professor emphasizes that AI does not make final judgments, stating that productivity has greatly increased as a result.
“Ultimately, our offers continue to evaluate an applicant’s eligibility. This method prevents anyone from being excluded”.
