Source: https://hbr.org/2021/09/ai-can-help-address-inequity-if-companies-earn-users-trust
The recent article of Harvard Business Review by Shunyuan Zhang, Kannan Srinivasan, Param Vir Singh, and Nitin Mehta highlights the challenges and opportunities that AI could offer to humanity. Similarly to what we discussed during the lecture, the authors emphasize on the numerous benefits that could be gained thanks to AI such as autonomous vehicles and automation within advertising decision, while at the same time pointing out some of the biggest challenges – such as the algorithmic bias.
However, the article builds a case for another crucial aspect in the success of AI, that reflects the challenge of rate of adaptability across diverse groups of people which should not be overlooked. The main argument in the paper is that the market conditions under which new algorithms are used should be carefully considered as this could determine the impact of the innovation as well as the benefits for the users.
The example of Airbnb launching AI algorithm-based smart pricing tool, used to automatically adjust listing’s daily price has shown exactly the importance of this process. Airbnb created an algorithm meant to help with determining the best price for property by analyzing a lot of information, and recently an increase of revenue by 8.6% of the hosts that adopted the algorithm was reported. Despite the results and the fact that the algorithm was perfectly race bling, the gap between balck and white hosts increased. The only reason for this was that black hosts were significantly less likely to adopt the algorithm in comparison to the while hosts.
Ultimately, this story shreds light on the importance of internalizing market conditions during algorithm deployment, meaning that differences in the adoption of AI algorithms across racial groups should be done by addressing the users concerns initially, to guarantee success and use by the targeted users.
Despite promising value to customers (bringing improved decision making) Airbnb’s algorithm ended up being unsuccessful. This sadly is not the only such story (e.g. reluctance in healthcare provided by AI, and Uniqueness neglect – Longoni, Chiara, Banezzi, 2019). Therefore, companies moving forward should focus on the key attributes to making AI beneficial – accountability, responsibility and transparency. The companies should consider how an algorithm could be perceived by targeted users and focus on building trust and helping users to understand what is the purpose of an algorithm and the way it works.
And how would you feel about using such an algorithm as a host in a platform such as Airbnb?
