University:
University College London
Course:
LAWS0022 | Law of Taxation
Academic year:

2024
Views:

71

Pages:

2
Author:

N. Betz

Algorithms In Immigration Administration Presented by: Joe Tomlinson Date: February 04, 2022 Overview - An algorithm is being used to sift couples into red and green risk categories - Red flagged couples face greater scrutiny and investigations, and potential sanctions - Minimal transparency around the eight criteria underpinning the algorithm The Bureau Of Investigative Journalism  Couples register intent to marry at a registrar's office.  If either is from outside the EEA + Switzerland, and does not have a visa or settled status, the proposed marriage is referred to the Home Secretary.  Information is fed into a triage system and flagged against 8 unknown "risk factors for sham marriage  Red-flagged couples are asked for more information, often with interviews & home visits.  If they do not comply, they will not be allowed to marry.  Immigration officers determine if it's a sham  If this decision goes against a couple, they can still marry.  But their immigration status is at risk and can lead to deportation. Sham MARRIAGES ALGORITHM 1. This is a typical example in UK immigration administration 2. No specific authority for use of automation - handled as an "operational issue" 3. Automation both replacing and supporting human officials 4. Limited transparency 5. Impact unknown and lack of official/public evaluation 6. Automation appears to use simple technology (less A.I., more if X then Y) Why Automate?  Government confidence with technology is growing (e.g. Government Digital Service)  Availability of technology has increased, including from the private sector  Lower costs and therefore savings for the taxpayer/redistribution to other policies  "Better” quality decisions (i.e. more formal consistency)  More data on decision-making (feedback loops)  Why not automate? Challenges For Lawyers 1. Automation is happening at a fast pace in government systems and it's having profound impacts 2. Multiple technologies involved (e.g. A.F.R, rule- based automation) and varied applications (e.g. decision-making, digital immigration status) 3. There is a sense we are in a period of experimentation where risks are being taken, with limited transparency, scrutiny, or mitigation 4. There is no single, overall answer on whether this is a good or bad development 5. Detailed systems analysis is needed, particularly in local settings and policy areas (e.g. housing) 6. Analysis needs to seek to understand impacts both top-down and bottom-up 7. There is a need for lawyers - working with other disciplines - to respond and innovate 8. A lack of certainty in the application of law, which is both a challenge and an invitation