AI and the Visa Process
Artificial Intelligence (AI) is making a big splash, particularly now with the roll out of systems like ChatGPT…whether it is really “AI” is another question but no doubt systems like this will only develop further and become better (or worse depending on how you view it) at what they do. They also have the potential to disrupt many different industries and the immigration process would not be immune. To illustrate the point, I asked ChatGPT a NZ immigration related question and here was the response:
“In New Zealand, a resident visa and a temporary visa are two different types of visas that have different purposes and requirements.
A resident visa is a type of visa that allows the holder to live and work in New Zealand indefinitely, as a permanent resident. It is designed for people who intend to stay in New Zealand long-term and who have the necessary skills, experience, or qualifications that are needed in the country. A resident visa can be obtained through different pathways, such as skilled migrant category, family sponsorship, or investment in the country.
On the other hand, a temporary visa is a type of visa that allows the holder to stay in New Zealand for a limited period, for a specific purpose. There are various types of temporary visas available, such as visitor visas, student visas, work visas, or partnership visas. Temporary visas are designed for people who intend to stay in New Zealand for a short period, such as for tourism, study, work, or to join their partner.
In summary, a resident visa is for those who want to make New Zealand their permanent home, while a temporary visa is for those who want to stay for a limited time for a specific purpose.”
As far as explaining the difference between a NZ Resident Visa and a Temporary one, this is pretty much spot on, although there are a couple of minor adjustments I would make. However it took less than four seconds for the system to generate this.
The use of systems like this isn’t new, although arguably the technology has rapidly improved and programs like ChatGPT and Bard are bringing those improvements in to the real world at a rapid pace. No longer are these systems and tools confined to the tech-savvy but they are now available to everyone and everywhere. It won’t be too long before various organisations begin to fold this technology in to their processes, albeit Government organisations might take a bit longer than most.
I remember attending an INZ seminar on what was back then (over a decade ago) their plans for a faster, more streamlined and more intelligent Visa processing system. I distinctly recall them referring to the use of AI to undertake risk profiling on various application categories to essentially automate the Visa process in certain situations. As of today, INZ have spent over $50 million on their new IT system and it is definitely artificial but far from intelligent. It is nothing more than a glorified website with a series of workflow steps engineered behind it - those steps trigger certain actions but are largely driven by data that is entered by the applicant or an immigration officer. There is no fear of INZ’s Visa system taking over the world any time soon.
That doesn’t mean however that it won’t happen and in fact back in 2010, the Government introduced a new section to the rules that referred to “automated decision making”. Over the course of 12 years that section was adjusted numerous times to reflect the increasing role that technology would play in determining Visas. The rules go on to essentially state that these systems can be considered to have the same processing authority (called warrants) as an immigration officer:
“…where an immigration instruction specifically mentions an immigration officer an automated electronic system may meet that instruction instead.”
These rules give INZ the ability to process Visa applications through a software system that has predetermined rules where that system assumes the role of an immigration officer in making the final decision - basically if you tick the right boxes, the computer will say yes, otherwise it will say no and your application will be declined.
In some cases this is actually pretty smart, particularly for low risk, high volume applications, where the criteria are relatively straight-forward and the risk factors can be identified and managed. However even in the simplest of cases, there are unintended consequences, for example how do you definie “low risk”? Is it based on where you come from or the passport you hold (those things can be quite different).
For anything more complicated, AI is likely to come unstuck. For example an applicant with a medical condition, where someone needs to review the history, treatment, impact, risk and costs against rules that are quite broad and by no means definitive, would leave even the most sophisticated AI scratching it’s digital head. Therein lies the danger however - there is a real pressure on Immigration New Zealand to deliver more Visas with less resources (they already run at a significant loss) and the use of systems to replace officers will become even more attractive over time. However the complexity of the Visa process means that if AI was adopted more broadly, we would be likely to see a growing number of incorrect decisions, leading to appeals, leading to more work and slower overall processing.
There is also a slightly less obvious move by INZ to remove the human element from processing, involving the use of generic email addresses instead of direct lines to case officers and correspondence being pushed through online portals rather than email or phone calls. Thankfully we still retain contacts throughout INZ where we can get in touch with the right people, however like most Government departments the move to a faceless, nameless interaction is inevitable - not one I would support however.
I was once asked by an industry colleague whether or not the move towards AI bothered me or made me wonder if I would need to pursue a new career - I didn’t hesitate in my response - absolutely not. The Visa process is more than just a form or a set of documents, it involves people and people are complicated. No matter how simple you think your application might be, it isn’t one that can be sifted through a series of clever algorithms. Every migrant brings their own challenges, issues and complexities to the process and it needs a human interaction to be managed properly.
Don’t get me wrong however, I do think there is a place for this technology and let’s face it, I dont think anyone can avoid it becoming a part of everyday life - so we might as well embrace the good parts. AI can be used in various steps within the Visa system to automate and speed-up processes that do chew up time - ultimately if that gives a human being more time to consider an application properly when making a decision, it can’t be a bad thing.
However we are miles away from any of that really impacting on the Visa process as it stands today. Considering the hefty price-tag for the system that INZ has now and the fact it is still full of holes - I don’t see Visas being processed by robots any time in the near future.