Historical education bolstered by AI-powered digital human

A digital human modelled and trained on the writings of Luigi Einaudi has been released by the Einaudi Foundation and Reply.

Using the capabilities and potential of generative AI to succinctly process and share information with end-users, the model is trained on archives and writings produced by Luigi Einaudi, the second president of Italy (serving from 1948 to 1955).

Digital services provider Reply joined forces with the Einaudi Foundation think tank, and civic organisation Fondazione Compagnia di San Paolo, to provide a generative AI model, based on GPT-3.5. This underpins an educational chatbot focused on Italian economic history.

Combining GenAI with hyperealistic 3D, the interactive model has the ability to answer the user’s questions in a manner which is consistent with the historical figure’s thought, overcoming geographical, physical and generational barriers.

“The project revolves around the liberal thought of Luigi Einaudi and it has certain characteristics, as it wants to make accessible all of Einaudi’s writings in the Foundation and on the web,” said Domenico Siniscalco, President of the Einaudi Foundation.

“This is a radical innovation that makes a leap forward in productivity and research possible, and also an expansion of the sources that can be used. I believe it will help keep the figure and thought of the President alive.

“In this sense, the project should not be used to look at the past, but to look forward, to the future.”

Francesco Profumo, President of the Compagnia di San Paolo Foundation, commented: “Faced with the multiple challenges that await us as a Foundation, we therefore confirm our commitment to providing answers to some of the social, environmental, technological, and economic challenges faced by philanthropic foundations.

“In particular, this project integrates the cultural heritage of Luigi Einaudi – economist, one of the masters of liberal thought, and the second president of the Italian Republic – with artificial intelligence and represents a step forward in the integration between cultural heritage and advanced technology.”

Transforming chatbot technology with GPT modelsTim Shepheard-Walwyn, technical director and associate partner at Sprint Reply, spoke to Information Age about how businesses can drive value from chatbot technology powered by GPT models.

Converging technologies

Machine Learning Reply built the generative conversational model specific to Luigi Einaudi’s writings, by applying a proprietary framework of algorithm training and results validation based on MFRAME Reply.

Meanwhile, Infinity Reply brought the digital human to life by replicating key elements of the historical figure’s appearance and iconic gestures thanks to 3D real-time technologies, and Bitmama Reply studied user experience and designed the communication campaign that will be presented online and in print in the coming months.

“Artificial intelligence is is rapidly permeating every aspect of our society, opening the door to new opportunities,” said Reply CEO, Tatiana Rizzante.

“Knowledge management is one of the areas to which Reply is paying particular attention. Managing knowledge with artificial intelligence means not only transforming the way data is accessed and information is extracted but also rethinking decision-making processes and the way people work.

“To support this change, we have conceived and developed MLFRAME Reply, a framework that integrates a proprietary methodology for database analysis with tools for creating conversational generative models applicable to specific domains of knowledge. This same framework today represents the intelligence component of Luigi Einaudi’s digital human, to which our team working on 3D real-time technologies has given a face and an image, this in essence being the creation of hyper-realistic digital humans.

“This synergy of skills and technologies has allowed us to extend the access to knowledge to a wider audience, creating an engaging connection between culture and people.”

The project makes a point of utilising up-to-date language, to ensure ease of use and value for educational institutions and students — the target user base for the stakeholders involved.

How it can be used

Through keyboard or voice input, students, specialists or other users can begin a conversation with the Einaudi digital human, around key topics related to the former President of the Italian Republic’s economic thought. These include:

  • banking;
  • biographical information;
  • competition;
  • inflation;
  • market and monopoly;
  • monetary and fiscal policy.

AI-led business processes – getting the balance right between business impact and staff satisfaction — How businesses can properly weigh up innovation and employee purpose, for long-term success.

Ground work

Fondazione Einaudi began this project in 2021, with the digitalisation of its archive of writings from Luigi Einaudi.

From here, the organisation opted to find a new way to more clearly present this historical information to the public using AI, through an interactive digital human, which led to the establishment of a fruitful and budding partnership with Reply.

Attention soon turned towards historical archives pertaining to economic policy and politics, more specifically a corpus consisting of 250,000 words taken from 30,000 pages and 15 original volumes and collections, available in digital format.

“What strikes me about AI is the speed of its spread and the depth of the recent debate,” Siniscalco added.

“In the spring and even in the summer months of this year, AI was a topic for the few. Today, it is on everyone’s lips and has captured the attention of all the media.”

It is expected that this project will be the first of many undertaken by this consortium of companies, with further pipelines to expand knowledge in the pipeline.

The “Pensiero Liberale, Dialogo Attuale” project is avilable and free to use on the Foundation Einaudi website.

This article was written as part of a content campaign with Reply.


84 questions to ask before training an AI modelAI experts and data scientists have published a checklist of questions developers should ask before embarking on training an AI model.

Avatar photo

Aaron Hurst

Aaron Hurst is Information Age's senior reporter, providing news and features around the hottest trends across the tech industry.