A significant limitation revolves around the data used to train our AI, requiring both accuracy and coverage. Accuracy refers to the quality of personality labels assigned to individuals, which must come from either self-aware individuals using scientifically validated quizzes or domain experts observing and inferring their likely type. Gathering and labeling tens of thousands of such data points pose a considerable challenge before our AI could be trained.
Coverage pertains to the diversity of individuals our AI can predict and accommodate, considering the various ethnicities and race groups worldwide. Given our AI's global reach, achieving high coverage was essential. The challenge arose due to the absence of a universally accepted standard for categorizing people, as this is influenced by many factors socially, culturally, historically, and also biological factors, rendering this a very complex and multifaceted issue. Additionally, there are also demographic factors such as gender and age that need consideration to ensure that our sample size in our training data could represent the broader global population accurately.
