What kind of a “Tinder for professional life” would you like to use?
The universal ambition of social networking is significantly hampered by the continuing physical separation and remote work. Without the serendipitous encounters at offices and events, how can we in the future best find new professional ties? How could IT perhaps help find better matches — new collaborators, more fit recruitments, and world-class teams?
Building on a recent Business Finland project, I want to point to our exciting article on computational social matching in professional life, published in the Communications of the ACM. The paper problematizes the topic from two central yet paradoxical viewpoints: human vs. computational decision-making.
On the one hand, computational support appears useful because the conventional human-driven matching is prone to biases in decision-making and the boundedly rational understanding of the breadth of alternatives. Human decision-making can result in homophily, the preference of like-minded others, and typically leans on the geographically limited pool of candidates. This may result in suboptimal collaboration and untapped co-creative potential, particularly in knowledge work and creative industries that demand cross-pollination of ideas and perspectives.
On the other hand, algorithmic matching is hardly a panacea to the challenge of finding “perfect matches”. It involves risks of strengthening the human biases due to biased training data for machine learning-based solutions, and of disregarding nuanced contextual details and human characteristics that would be relevant to the matchmaking decision. For example, the notion of a perfect match does not generalize across individuals, and the same well-regarded individuals cannot practically be recommended to everybody (that is, the Matthew effect). The article reminds us that directly applying the prevailing algorithmic models can introduce new risks with detrimental effects on the performance, wellbeing, and collaboration practices in work life.