Skip to main content Skip to navigation

Warwick Manufacturing Group

CADE Workshop ‘24: Generative AI in the Digital Economy

CADE Workshop ‘24: Generative AI in the Digital Economy

June 25, 2024 | 09:00-12:00 (Italy Local Time)

Organisation Committee:

  • Alpay Sabuncuoglu, The Alan Turing Institute
  • Carsten Maple, University of Warwick, The Alan Turing Institute
  • Andrew Elliott, University of Glasgow, The Alan Turing Institute
  • Marie Briere, Amundi Investment Institute, Université Libre de Bruxelles & Paris Dauphine PSL University
  • Michela Iezzi, Banca d’Italia

Background:

The digital economy refers to economic activities that rely on digital technologies, such as the internet, mobile devices, and digital platforms, to create, distribute, and consume goods and services. It encompasses various sectors including e-commerce, online banking, digital marketing, and telecommuting, enabling businesses and individuals to conduct transactions, communication, and commerce electronically, leading to increased efficiency, innovation, and globalization of markets.

Machine learning (ML) systems have become integral components of the digital economy and have been revolutionising various tasks such as financial modelling, security analysis, fraud detection and customer interaction. From linear models utilised in the scoring systems to neural nets in biometrics for customer authentication, these different ML applications powered an extensive array of indispensable applications in the digital economy.

In the past decade, generative AI has become an interest to many considering their advanced capabilities in processing large volumes of data and generating plausible outputs. Their ability to demonstrate success in these different tasks can be bound to two properties: The scale of the training data and the size of the model architecture.

Based on the output modalities, we can categorise generative AI models as language models, image models, video models, multi-modal models, etc. While language models excel in text-based tasks such as language translation, summarization, and information processing, an image model can generate thousands of synthetic face data which can help us to develop face biometric systems in a privacy-preserving way. The multi-modal production capabilities of Generative AI also supported practitioners in achieving privacy, security and productivity in application development by demonstrating significant improvements in synthetic data production, unstructured data understanding, and multi-modal data interpretation.

The opportunities presented by generative AI and LLMs are vast. Currently, businesses of varied sizes and sectors have started using these models to power up their products before comprehensively understanding the nature of data and models. They leverage these technologies for analysing complex unstructured data, generating insights, customer engagement, or automation of repetitive tasks.

However, alongside these opportunities, challenges persist. In this dynamic and sensitive environment, proactively incorporating security, privacy and fairness considerations is crucial.​​ For example, biases inherent in training data can perpetuate societal inequalities and undermine the fairness of AI systems. Additionally, the resource-intensive nature of training and deploying advanced AI models poses scalability issues for smaller organizations. Further, their data and model size and black-box structure raised concerns over the safe integration of these applications. Although it is a long-time interest of NLP researchers to develop robust, secure, and fair language models, we are still struggling to achieve these characteristics that build up safety. This is partly due to the complex nature of these concepts, and the decentralized efforts of design, development and assessment groups.

Purpose:

On the 27th of March, The Alan Turing Institute (the Turing) and HSBC with colleagues from various financial institutions including high-street banks, accounting, consulting and insurance companies, and other financial professionals published a report on the integration of Large Language Models (LLMs) in financial services (see the report hereLink opens in a new window). In this report, researchers delved into the potential opportunities and challenges of integrating LLMs into financial services. A key recommendation for the academic community was defining the research question with a focus on specific use cases, considering the current regulations and internal managerial frameworks to yield high-impact research outputs.

Following this key recommendation, the researchers from the Turing and University Warwick invite you to submit your original research outputs, work-in-progress reports, and opinion pieces to our half-day workshop on “Generative AI in the Digital Economy.” In this workshop, we aim to establish a collective understanding of “trustworthiness,” specifically security, privacy, and fairness, in the development and use of generative AI in digital economy-related services. These services include but are not limited to:

  • Financial insight generation (e.g. report generation, forecasting),
  • Money economics-related services (e.g. treasury operations, credit lending),
  • Financial services safety and security (e.g. monitoring threats, fraud detection),
  • Customer interaction and communication (e.g. chatbots, financial literacy),
  • Digital Identity and Know-Your-Customer (KYC) solutions (e.g. biometrics, anomaly detection)

During the first half of the workshop, we will engage with panel discussions and brief lightning presentations. Following this, the latter half will be dedicated to exploring the critical issues highlighted in the report and presentations. Together, we aim to chart future research avenues for generative AI in the digital economy.

Agenda:

The workshop will have a blended format, with participants being able to present and attend in person or virtually.

Date and Time: June 25, 2024. Italy Local Time, UTC+2, CEST

  • 8:30-9:00 – Registration, Coffee and Networking
  • 9:00-9:10 – Welcome
  • 9:10-9:40 – LLMs in Finance Panel
  • 9:40-10:30 – Lightning Presentations
  • 10:30-10:45 – Break
  • 10:45-12:00 – Discussion and Horizon Scanning

Workshop Location:

The workshop is co-located with CADE 2024, which will be held at the Fondazione Levi, overlooking the Grand Canal in Venice.

More information about the location can be found here- https://www.fondazionelevi.it

Submitting Papers:

We invite submissions encompassing academic research, industry practices, critical reviews, and diverse viewpoints. We encourage academics, practitioners, and policymakers to contribute short papers that stimulate discussion in the workshop.

Format: The submission should not exceed a maximum of two pages, excluding references. While a specific format is not mandated, we strongly encourage submissions to be visually appealing and easily readable. Submissions should not be anonymized, and at least one corresponding author should attend the workshop.

You can submit your paper through our Zenodo Community Page:

https://zenodo.org/communities/genai-digital-economy

If it is your first time submitting a paper to an open-source publishing community, the following help guide explains how to create a submission and submit it for review: https://help.zenodo.org/docs/share/submit-for-review/

Important Dates:

All the dates are AoE.

  • Apr 26, 2024: Paper submissions
  • May 10, 2024: Reviews sent to authors
  • May 24, 2024: Rebuttal period ends
  • June 3, 2024: Notification of acceptance/reject
  • June 14, 2024: Camera-ready submission

Registration:

TBD.