Bad Robots: Behind the Tinder Algorithm – the Unknown Truth about Tinder

Bad Robot Outcome: Bias Tinder algorithm learns to make predictive future projection decisions based on past habits for future matches. It also poses a concern about how the algorithm will objectively calculate personal desirability.

They say, “The best predictor of future behaviour is past behaviour” (Mark Twain). Artificial intelligence systems have continually been applied to every sector to improve how things were done previously. However, these new innovations and applications raise ethical questions, especially on whether they observe and follow ethical principles set out by governments and relevant standard bodies in the development of smart systems and AI. In the case below we will review the ethical concerns around the tinder algorithm for its dating site. Tinder is a global online dating and networking application that allows users to anonymously swipe to like or dislike other profiles based on their photos, a small bio, and common interests. Once two users have “matched”, they can then exchange messages. The Tinder algorithm is an AI algorithm developed and implemented whose main purpose is matchmaking between its users. Experts have questioned this algorithm concerning its operation, and ethical principles since reports of individual complaints about the algorithm are somewhat biased.

The Story

The algorithm works based on learning users’ past habits. For instance, if an individual had, several Caucasian or blonde matches in the past, the algorithm is then designed to suggest only Caucasian or blonde as the perfect match for the user in future matches. This raises ethical concerns and poses a threat to human interaction and social life in dating sites since it reinforces societal norms.
If a person made discriminatory decisions, the algorithm would continue to reinforce those decisions based on its prediction trajectory. It has no mechanism to suggest something different and thus reinforces the status quo that could be damaging to an individual or society. The algorithm is also programmed to convey the “algorithm of desire”. This means that tinder users’ score will be based on how often a user logs in. A lower score means that one cannot swipe through top-rated tinder users with a great score. This desirability game on the tinder platform operates on factors linked with personal preference but are not universal to all users; hence not objective rather than subjective. It thus poses a concern about how the algorithm will objectively calculate personal desirability. Though rejections were removed from the Tinder platform, the altered rejection process is a bit worrying since the number of left swipes is kept hidden from its users, as are right swipes. The algorithm can deny someone a match or matches by keeping off the likes from someone through this process.
This is thought to have been put in place to slow down the upper percentages of most desirable users and hence to make their profile less visible to another tinder user to promote people with lower rankings to get a match.

Our View

It is clear as we have seen that the Tinder algorithm is still a work in progress as it develops, grows, and continually keeps fine-tuning its algorithm to meet the needs for their dating site. However, there is still much to be done regarding the arising ethical concerns of the algorithm in place and in the world of dating apps. The algorithm should allow for independent decisions and not upon past information that could be biased and potentially discriminatory to certain Tinder community members. Most importantly it must uphold the equality of all Tinder users regardless of their race, class or status quo. Otherwise, the algorithm will end up doing more harm than good in the long term. It infringes on Fairness and Contestability (Australian Gov’t AI ethics Principles). We hope to see bias eradicated from dating apps where it’s an even and equal playing field for all its members irrespective of input.


Lawrence, s. (2019, May 18). The Tinder algorithm explained. Retrieved from VOX:

Rolle, M. (2019, 02, 25). The biases we feed to Tinder algorithms. Retrieved from


Sarah Klain

Written by:

Sarah Klain