The Vatican’s media outlet, Vatican News, reported that the Asian synodal continental assembly in Bangkok had used artificial intelligence (AI) to help draft a final document, sparking an online debate.
The assembly’s participants represented 29 countries, and Fr Clarence Devadass, a discernment and drafting team member, presented the updated document.
Fr Devadass explained the process: “From the data received, the responses were inputted into an AI software, and a command was given like, ‘From the response below, highlight the common themes’ or ‘Which are the responses that are specific only to one particular group?'”
Once the data was sifted through, the drafting committee rechecked it to ensure accuracy.
Some online conversations regarding the report raised misconceptions about the process. While “artificial intelligence” may sound like science fiction, it describes computer systems designed to do tasks typically associated with humans, such as identifying what a picture depicts, transcribing spoken communication, or discerning a meaning in written text. The training process for AI models is always human-assisted.
Fr Devadass said the AI system effectively sorted data and picked up on keywords. Still, it needed to be monitored by human resources to ensure the accuracy of the data. But, he added that what would have taken a couple of hours was done in a few minutes.
The use of AI in drafting documents raised concerns about the extent of the system’s role in the final product. Still, the Asian assembly’s experience demonstrated the effectiveness of the technology as a tool for sorting and processing data.
Use of AI “not new” and “helpful”
“The use of artificial intelligence and digital technologies in such processes is not new,” said Dr Paolo Benanti, a theologian at the Pontifical Gregorian University in Rome. “Indeed, it can be very helpful, especially when dealing with a large amount of data. However, it is essential to be very careful in the use of these tools, and human control and guidance should always be present.”
The report on the use of AI at the Asian synod meeting highlighted the increasing use of natural language processing. This is AI technology that categorises written text in terms of content and tone. Natural language models are trained on large sets of written text. Then human trainers identify specific topics and examples of tone. For example, this technology is used in customer service requests to identify angry or sarcastic prose.
In the Asian synodal document, the AI system was used to identify common topics and flag outliers. While the process may not have been perfect, it demonstrated the potential of AI to streamline and assist in the drafting process while requiring human oversight to ensure accuracy and integrity.
Sources