Rep Elections: Candidate suspected of using AI and switching pronouns 

A candidate running for seven positions in the Rep Elections is believed to have used AI in his manifestos and switched pronouns to qualify for positive action quotas.

Lola Davies

Serial candidate Yixiao Huang is simultaneously running for Accommodation and Housing Officer, NUS Conference Delegate, Societies Rep for Student Media, Student Trustee, Undergraduate Faculty Rep for Social and Historical Sciences, Disabled Students’ Network Social Secretary, and People of Colour Network Social Secretary.

The candidate-of-seven appears to be exploiting the positive action quota for women and non-binary people when he opted to use the neopronouns xe/xem/xyr in his NUS rep nomination, despite using he/him/his on his LinkedIn profile and in his six other nominations.

Additionally, an analysis by The Cheese Grater can reveal that it is very likely Huang has written his manifestos with the help of artificial intelligence (AI).

The Cheese Grater analysed each of Huang’s manifestos at the Rep Elections using two different AI detection tools, Grammarly Pro and Quillbot. 

In the races for NUS delegate, Social and Historical Sciences UG Rep, and Student Media Rep, Huang is the only nominee whose manifestos are likely to be at least partly AI-generated, as his co-nominees’ manifestos all return a 0% chance of AI generation. 

In the remaining four races, the majority of his co-nominees’ manifestos returned either 0% likelihood scores or a score within the margin of error. 

The table below shows the scores provided by the two AI detection tools for each role Huang is standing for. 

The discrepancy between the scores given for the same text can be explained by the fact that AI detection is underdeveloped, and a margin of error must be accounted for.

While AI detection software isn’t foolproof, Huang’s likely AI use is discernible to the naked eye.

ChatGPT has the tendency to use repeated language in its responses like ‘delve’, ‘embark’, ‘enhance’ and ‘foster’. The list goes on, and if you’ve ever used the tool in any capacity, you know exactly the type of language we mean.

Incidentally, Huang’s manifestos are full of this language, including the use of ‘foster’ a whopping 11 times across his seven nominations.

Readers should know that using AI to mass-produce election pitches isn’t strictly against Union rules, nor is being inconsistent about your pronouns.

Neither is Huang alone in (probably) using AI in his manifestos, although he is certainly one of the most prolific candidates to have done so.

It is a strategy that might very well see him land the position of Social Secretary Rep for the Disabled Students’ Network, for which he is the only standing candidate, apart from re-open nominations.

But this is only the case because 60% of roles at this election are either vacant or uncontested, meaning there is no genuine competition to deter candidates like Huang, who represent the symptom, not the cause of student apathy.

Until we start taking the Rep Elections seriously, ‘Grifting is back on the menu’, as one candidate for Student Trustee said.

Yixiao Huang has denied all allegations and responded with the following comment:

‘I should claim that your accusations are false and will cause a loss in my reputation. So I request you to stop this making-up behaviour and write nothing but truth.

‘I won’t give permission to The Cheese Grater Magazine for wirting [sic] a story about me.’

Unfortunately, that is not how libel works.

Students’ Union UCL responded with the following:

‘The Representative Election is run in accordance with our Election Byelaws and overseen by an external Returning Officer.

‘Any student can report concerns to the Returning Officer via the Deputy Returning Officer who will investigate.

‘More information can be found here: https://studentsunionucl.org/election-rules.’

This article featured in the Digestive – Special Issue