Skip to content.

Introducing the Use of Automation Tools in Canadian Immigration Processing

Inland Immigration processing time has been challenging for years but it reached an all time high after the onset of the pandemic. For at least the last two years work permit extensions and inland work permit applications have been taking 5-6 months to process due to an unprecedented volume of applications and labour shortages.  Immigration, Refugee and Citizenship Canada (IRCC), has acknowledged these issues and have started leveraging the use of automation tools to alleviate the current strain on the system.  

Delays in processing work permit extension applications are problematic when applicants are still awaiting a decision on their application, but their existing work permit has already expired putting them into maintained status (previously known as implied status)Although you can continue to work while on maintained status, not having a valid work permit can cause many issues including the following:

  • Inability to extend Provincial Health Care coverage;
  • Confusion due to expiring Social Insurance Numbers (SIN);
  • Inability to renew credit cards;
  • Inability to gain loans or mortgages if needing to buy a car or a house;
  • Inability to travel outside of Canada for personal or business reasons; and
  • Prevents employment mobility with the same or another employers.

Although there are ways to apply for a new work permit in extenuating circumstances, this still creates a great deal of frustration and hardship for both applicants and employers.

IRCC has been working on aggressive hiring measures to help alleviate the processing backlog however, they are also suffering from low unemployment rates and employee retention issues, just like many other employers.  It also takes time to train and develop new staff.  The result being that despite IRCC’s best efforts, processing times stayed steady at approximately 142-147 days for quite some time.

Then…. on September 26, IRCC announced that they are introducing automation tools to assist in processing inland work permit extension applications as well as post graduate work permits (PGWP).  For now, these tools are only being used by the inland processing centre and we can already report a decrease in processing times to as low as 128 days.  Although this is only about a two-week difference, it is the largest improvement in processing times in a year.

IRCC has been working on this initiative for several years and was already engaging in consultations on the use of AI in immigration processing back in 2019.  The consultation process in which system bias, privacy, consistency in decision making, potential manipulation of data and applicant vulnerabilities were all raised, seems to have been an effective process as all these issues were taken into consideration when these tools were designed. 

The goals of these process improvements are simply to increase processing efficiency and to decrease processing time and are not meant to replace the discretion of a reviewing officer.  The current tools have two functions:

  1. To triage cases into processing buckets based on complexity and to assign cases to an appropriate officer; and
  2. For straightforward cases, to confirm eligibility for the program under request.

Applications are then forwarded to an officer to assess for admissibility issues.  A final decision is rendered, and the decision is finalized.

Although the automation tools can approve an application, they are unable to deny an application and a negative determination on an application by the tools must go to an officer for review and a manual decision.  If the application is ultimately refused, the refusal letter should include an outline of the use of the automation tools in the determination of the application along with written reasons for the refusal.  

One of the key questions about this movement toward automation is how will the program ensure that it maintains it’s integrity as well as an absence of systemic bias.  Here is what IRCC has advised about their processes in this regard:

  • All tools have undergone an Algorithmic Impact Assessment (AIA) before being launched to review potential discriminatory impacts, and privacy and security elements built into the design;
  • The tools themselves and the rules used to guide the tools have been created by IRCC staff and have not themselves been developed by AI or other machine learning tools;
  • The rules used to develop the tools are based only on data elements with a clear link to legislative and regulatory requirements;
  • The system, and the rules used to drive the system, and final decisions will be routinely reviewed to ensure results are consistent and as intended, as well as reflective of changes in law and policy;
  • Applications can only be approved by the automation tools and negative decisions must always be reviewed by an officer, a reviewing officer always has the ability to overturn a tool-based determination; and
  • Officers are trained not to let how the files are triaged influence their final decisions.

At this time there are already other programs that are utilizing these or similar tools such as applications for temporary resident visas (TRVs), visitor record requests, study permits and study permit extensions, privately sponsored refugee applications, spousal sponsorships, TR to PR applications, and more programs to be launched shortly.   Visitor record extension applications which were taking over 200 days to process six months ago are already down to 49 days as of the writing of this article.

So, what does the future hold?  We will likely see faster processing times in the near future as well as decisions that are much more technically and objectively based.  However, although the new technology is exceptionally helpful, we also anticipate seeing a significant learning curve as the system goes live and expect unexpected complications in processing and an increased number of documentary errors that will need o be corrected during the implementation period.  So, although the long-term prospects with these automation tools are optimistic, there may be some bumps along the way.

Authors