AI's Attempt to Predict Papal Election Falls Short, Highlighting Challenges in Forecasting Unique Electoral Processes

AI's Papal Prediction Misses the Mark

In a recent development that has caught the attention of both the tech and religious communities, an artificial intelligence (AI) algorithm's attempt to predict the next pope has fallen short. The Vatican announced that U.S. Cardinal Robert Prevost would be the next pope, a choice that was not on the AI's shortlist 1

1

The AI Model and Its Prediction

A team of researchers, including Eugenio Valdano from INSERM and Michele Re Fiorentin from the Polytechnic University of Turin, developed a machine learning algorithm to predict the outcome of the papal conclave. The model analyzed the ideological positions of 133 high-ranking Catholic Church officials on four key topics: attitudes toward same-sex couples, international migration and poverty, interfaith dialogues, and synodality 1

1

The AI model predicted that Italian Cardinal Pietro Parolin would be the next head of the Catholic Church. Other candidates highlighted by the model included South African Archbishop Stephen Brislin and Filipino Cardinal Luis Antonio Tagle 1

1

Challenges in Predicting Papal Elections

The researchers acknowledged the unique challenges posed by papal conclaves. Unlike political elections, conclaves generate very little data, with the process remaining largely secretive. The dynamics of an election involving a small group of people choosing from among themselves differ significantly from large-scale political elections 1

1

Related Stories

AI Models and Prediction Markets Diverge on Next Pope Selection

Pope Leo XIV Rejects AI Avatar, Warns of Technology's Impact on Humanity

Pope Leo XIV Addresses AI as a Major Challenge, Echoing Historical Church Response to Industrial Revolution

Anthropic's Chris Olah tells Vatican that AI development cannot be left to Big Tech alone

AI Chatbots' Approach to Papal Predictions

In a separate experiment, various AI chatbots were asked about the potential next pope. Interestingly, these advanced systems showed a reluctance to make definitive predictions:

1. OpenAI's ChatGPT acknowledged the lack of clear consensus and listed potential frontrunners without endorsing any specific candidate 2

   2

   .

2. Grok, developed by X (formerly Twitter), provided a detailed analysis of factors influencing the conclave but refrained from naming a preferred choice 2

   2

   .

3. Claude, Anthropic's chatbot, adopted a reverent tone and focused on challenges facing the Church rather than endorsing candidates 2

   2

   .

Implications for AI in Electoral Predictions

While the AI's prediction for the papal election was incorrect, experts believe that the approach could still be valuable for other types of electoral contests. Data scientist Rohitash Chandra from the University of New South Wales Sydney suggested that the method of categorizing candidates' positions on specific topics could be applied to local elections 1

1

However, the limitations of the model were also highlighted. Factors such as geography, popularity, and private discussions among electors during the conclave were not fully accounted for in the AI's predictions 1

1

This event serves as a reminder of the complexities involved in predicting unique electoral processes and the current limitations of AI in fully capturing the nuances of human decision-making in such contexts.