In a significant stride for international collaboration in AI ethics, José Antonio Siqueira de Cerqueira, a Doctoral Researcher at the GPT Lab, visited the University of Copenhagen (UCPH), Department of Computer Science, from April 23 to April 27, 2024. His visit, hosted by Dr. Ville Vakkuri of the Human-Centred Computing (HCC) section, focused on advancing our understanding of trust in large language models (LLMs) through a novel multi-agent approach.
Exploring Trust in AI
José Antonio’s research at the GPT Lab centers on the ethical implications of artificial intelligence. His current work seeks to address the challenge of developing ethically aligned AI-based systems. He argues that an enhanced LLM-based system can significantly contribute to this goal by providing practical, actionable guidance to developers, who often struggle with the abstract nature of ethical guidelines.
During his time at UCPH, José Antonio explored how a multi-agent approach, utilizing LLMs, can foster trust in AI systems. By simulating interactions among various AI agents, researchers can gain insights into the ethical dimensions of AI behavior, potentially leading to more trustworthy and ethically sound AI applications.
Highlight of the Visit: Lunch Talk on Trust in AI
A key event during José Antonio’s visit was his lunch talk on April 26, 2024, held within the HCC section. The talk, titled “Trust in AI: LLM-based Multi-Agent Systems & Finnish AI Presidential Debate,” provided a comprehensive overview of his PhD research at Tampere University and the innovative Finnish AI Presidential Debate project.
Talk Highlights:
- Topic: Trust in AI: LLM-based Multi-Agent Systems & Finnish AI Presidential Debate
- Date: April 26, 2024
- Venue: Human-Centred Computing section, UCPH
José Antonio’s talk delved into the complexities of building trust in AI systems through LLM-based multi-agent models. He emphasized the need for practical examples and robust governance to guide developers in creating ethically aligned AI systems.
The Finnish AI Presidential Debate: A Synthetica Project Initiative
In addition to his primary research, José Antonio discussed the Finnish AI Presidential Debate, a groundbreaking initiative under the Synthetained attention, as highlighted in this article by YLE.
Synthetica Project page: https://www.tuni.fi/en/research/synthetica.
Looking Ahead
José Antonio Siqueira de Cerqueira’s visit to the University of Copenhagen has further strengthened the ties between the GPT Lab and international AI research communities. His contributions underscore the importance of collaborative efforts in addressing the ethical challenges posed by advanced AI systems.
As we reflect on this fruitful exchange, we anticipate future collaborations that will continue to push the boundaries of AI research, ensuring that ethical considerations remain at the forefront of technological advancement.
Fun fact: this blog post was assisted by an AI. Here’s to the wonders of technology!