The Home Office is using an algorithm to determine whether a marriage should be investigated as a ‘sham’, according to documents obtained by the Public Law Project under the Freedom of Information Act.
PLP is concerned that this algorithm may be flawed and discriminatory because some nationalities seem more likely to be targeted for investigation than others.
Get in touch
Our Research Team is keen to hear from anyone who feels they have been unfairly targeted by a sham marriage investigation, and from any organisations working in this area. You can contact Jack Maxwell, one of our Research Fellows, at j.maxwell@publiclawproject.org.uk.
Background
The UK has a detailed legal framework for targeting ‘sham marriages’: where couples get married to avoid immigration law, rather than because they have a genuine relationship. As part of this framework, registrars are required to refer proposed marriages to the Home Office if either or both of the parties is subject to immigration control.
Since at least April 2019, the Home Office has used an algorithm to triage these referrals. The algorithm gives each couple a ‘Red’ or ‘Green’ rating. ‘Red’ means that the Home Office should investigate the couple ‘to rule out or identify sham activity’, while ‘Green’ means that an investigation is not warranted. Media reports indicate that sham marriage investigations can be highly invasive.
The core elements of the Home Office’s algorithm are set out below:
PLP is concerned that this algorithm may be flawed and discriminatory. In particular, the Home Office’s documents show that some nationalities, including Bulgarian, Greek, Romanian and Albanian people, have their marriages rated ‘Red’ at a much higher rate than others. The Home Office has – so far – refused to disclose all of the ‘risk factors’ used by the algorithm to rate a case. The risks of discrimination in algorithmic decision-making are well-known. In August last year, the Home Office scrapped an algorithm it used to help decide visa applications, in the face of allegations that it was racially discriminatory.
PLP’s Research Team continues to investigate this algorithm, as part of our ongoing work on tracking automated government.