Further analysis & Evaluation.
Match Pro proved not to be the best example of AI and tech driven dating platforms given its revenue targets were clearly a dream with no evidence or documentation of any sort of product, not even prototypes, it was hard to analyse a passage of text that was completely immaterial.
However the actions of current platforms and trends in emerging services also suggest this is where we are heading so I took more of a holistic approach by analysing the entire field and itâs trend towards AI and data driven match making.
Collective analysis of Artificial intelligence in matchmaking.Â
-There is nothing particularly new about machine algorithms in dating services. What is new is the vast amount of data these algorithms now have access to (our online footprint)
-MatchProâs selling point was data driven results based off psychological and graphological analysis however currently dating apps are able to provide recommendations based off of the analysis of user history which you could argue is a stronger indicator. From age, gender, location, and sexual preferences, to online purchasing history and even Spotify playlists
-Iâve noticed two distinct approaches, the first being relying on data that already exists e.g social media & the online footprint and the scientific approach being extensive surveys and tests. These platform tend to target different demographics. Platforms like eHarmony and OKCUPID take this approach to data collection however people have voiced their concern over the opportunity for users to miss represent themselves. Amy Webb a journalist and professor of strategic insight at the stern school of business documented her ability to out smart a match making algorithm by catering her profile to the man she hoped to meet.
A quote from the CEO of tinder Sean Rad speculating on the state of the platform in 5 years time. âThe Tinder voice might pop up and say, âThereâs someone down the street that we think youâre going to be attracted to, and sheâs also attracted to you, and guess what, sheâs free tomorrow night! And we know you both like this indie band, and itâs playing, so would you like us to buy your tickets?â
Platforms like AIMM (1) seem the be smarter in the way they go about collecting data to fuel their algorithms. The application is voice controlled which allows the AI to pick up sentiment in the voice of the user, whilst platform like Loveflutter (2) analyse users twitter history to pick up sentiments and make assumptions about their personality.
âtodayâs dating sites are only as good as the data theyâre given.â
Value proposition of AI and big data in dating platforms:
-Vast pool of potential matches, instant gratification, convenience
-âAlgorithms can end up knowing a person better than friends, family or even themselves, and thatâs revolutionising matchmakingâ, says Michal Kosinski, a computational psychologist and assistant professor at Stanford Universityâs Graduate School of Business.
- âAlgorithms can learn from experiences of billions of others, while a typical person can only learn from their own experience and the experience of a relatively small number of friends.â refers to the importance of Big Data.
-Signals that people may miss but AI can catch include the sentiment in communication, response times, and length of profiles.
-With such a large and accessible candidate pool there are no missed opportunities.
It seems as though a lot of common dating speed bumps can be avoided with the due-diligence relatively simple AI can provide but do these algorithms need to be inherently broken? Otherwise successful users will not return to the platform.Â
Evaluation.
Questions: How does the quality of interaction differ when first meetings are based upon a pool of compatible candidates and alternatively by chance? How does the context of the situation (expectations of a date) alter the course of the interaction?
Is it ultimately more rewarding to meet someone by chance or serendipity or through a controlled and monitored pool of candidates?
These platforms market themselves on the potential for love and this is where the criteria for success becomes pretty muddy. When looking simply at introductions it is a bit easier to process. Instead of a date I chose a single test  subject (one of my flatmates) to choose a particular activity to complete with their date (another of my flatmates) The subject chose a game of call of duty.
I choose to base the meeting around an activity as these were people who already new each other and I thought this provided some criteria to which they could make judgements about each other. I put no restriction on the type of activity. The rest of my subjects (my remaining flatmates) and my initial subject wrote on a piece of paper what their chosen activity would have been and two key skills that related to their chosen activity. My attempt to simulate Artificial intelligence was to use actual intelligence in the form of two other friends. In this test subjects would be paired to complete the chosen activity first by the âAIâ and next at random, they would then reflect on the experience and rate the quality of interaction based on whether they would like to engage in that particular activity with their partner again. The two match makers who played the role of a machine algorithm were told to make a decision purely on the information that was provided by each subject, this consisted of three key pieces of information. The match makers selected two people for the subject to complete the selected activity. Then a partner was chosen at random.
In hindsight this was a bit of a stretch. It is extremely difficult to replicate the chemistry of potential partners meeting as well as triggering the same response as people meeting for the first time. Although this was an interesting test Iâm hesitant to make any inferences about the effectiveness of artificial intelligence in dating.
I did notice that the context of the situation had an effect on their attitude, when their partner was chosen from a pool they became fixated on executing the task whereas when the partner was chosen at random was more concerned with having a laugh and looked to be enjoying the process more. Being put in a situation with someone who is said to be a match tended to bring out a more assessment oriented attitude. subjects were actively try to find value in the other person. As soon as there was a slight amount of disinterest they would become totally disinvested with the mindset there would likely be another better partner. When the activity partner was chosen at random they tended to be a lot more open mindedÂ
Iâd like to know if this same attitude is present during first dates? Are daters concerned primarily with the conversion rate of meetings or do they enjoy the process?
(1)Â https://aimm.online
(2)Â https://www.loveflutter.com












