How can we make the most effective possible use of enormous language models for a better and more inclusive society?

What do you do when you do not know a term like “LLM”? You most likely quickly google it or ask your team. We use the knowledge of groups, generally known as collective intelligence, as a matter in fact in on a regular basis life. By combining individual skills and knowledge, our collective intelligence can achieve outcomes that exceed the capabilities of any individual alone, even experts. This collective intelligence drives the success of every kind of groups, from small teams within the workplace to massive online communities like Wikipedia and even societies at large.

LLMs are artificial intelligence (AI) systems that analyze and generate text using large datasets and deep learning techniques. The brand new article explains how LLMs can enhance collective intelligence and discusses their potential impact on teams and society. “As large language models increasingly shape the data and decision-making landscape, it’s crucial to strike a balance between harnessing their potential and safeguarding against risks. Our article details ways through which human collective intelligence may be enhanced by LLMs, and the varied harms which might be also possible,” says Ralph Hertwig, co-author of the article and Director on the Max Planck Institute for Human Development, Berlin.

Among the many potential advantages identified by the researchers is that LLMs can significantly increase accessibility in collective processes. They break down barriers through translation services and writing assistance, for instance, allowing people from different backgrounds to participate equally in discussions. Moreover, LLMs can speed up idea generation or support opinion-forming processes by, for instance, bringing helpful information into discussions, summarizing different opinions, and finding consensus.

Yet the usage of LLMs also carries significant risks. For instance, they may undermine people’s motivation to contribute to collective knowledge commons like Wikipedia and Stack Overflow. If users increasingly depend on proprietary models, the openness and variety of the knowledge landscape could also be endangered. One other issue is the chance of false consensus and pluralistic ignorance, where there may be a mistaken belief that almost all accepts a norm. “Since LLMs learn from information available online, there may be a risk that minority viewpoints are unrepresented in LLM-generated responses. This could create a false sense of agreement and marginalize some perspectives,” points out Jason Burton, lead creator of the study and assistant professor at Copenhagen Business School and associate research scientist on the MPIB.

“The worth of this text is that it demonstrates why we want to think proactively about how LLMs are changing the web information environment and, in turn, our collective intelligence — for higher and worse,” summarizes co-author Joshua Becker, assistant professor at University College London. The authors call for greater transparency in creating LLMs, including disclosure of coaching data sources, and suggest that LLM developers needs to be subject to external audits and monitoring. This might allow for a greater understanding of how LLMs are literally being developed and mitigate adversarial developments.

As well as, the article offers compact information boxes on topics related to LLMs, including the role of collective intelligence within the training of LLMs. Here, the authors reflect on the role of humans in developing LLMs, including the best way to address goals akin to diverse representation. Two information boxes with a give attention to research outline how LLMs may be used to simulate human collective intelligence, and discover open research questions, like the best way to avoid homogenization of data and the way credit and accountability needs to be apportioned when collective outcomes are co-created with LLMs.

Key Points:

• LLMs are changing how people seek for, use, and communicate information, which may affect the collective intelligence of teams and society at large.
• LLMs offer recent opportunities for collective intelligence, akin to support for deliberative, opinion-forming processes, but in addition pose risks, akin to endangering the range of the data landscape.
• If LLMs are to support quite than undermine collective intelligence, the technical details of the models should be disclosed, and monitoring mechanisms should be implemented.

Participating institutes

• Department of Digitalization, Copenhagen Business School, Frederiksberg, DK
• Center for Adaptive Rationality, Max Planck Institute for Human Development, Berlin, DE
• Center for Humans and Machines, Max Planck Institute for Human Development, Berlin, DE
• Humboldt-Universität zu Berlin, Department of Psychology, Berlin, DE
• Center for Cognitive and Decision Sciences, University of Basel, Basel, CH
• Google DeepMind, London, UK
• UCL School of Management, London, UK
• Centre for Collective Intelligence Design, Nesta, London, UK
• Bonn-Aachen International Center for Information Technology, University of Bonn, Bonn, DE
• Lamarr Institute for Machine Learning and Artificial Intelligence, Bonn, DE
• Collective Intelligence Project, San Francisco, CA, USA
• Center for Information Technology Policy, Princeton University, Princeton, NJ, USA
• Department of Computer Science, Princeton University, Princeton, NJ, USA
• School of Sociology, University College Dublin, Dublin, IE
• Geary Institute for Public Policy, University College Dublin, Dublin, IE
• Sloan School of Management, Massachusetts Institute of Technology, Cambridge, MA, USA
• Department of Psychological Sciences, Birkbeck, University of London, London, UK
• Science of Intelligence Excellence Cluster, Technische Universität Berlin, Berlin, DE
• School of Information and Communication, Insight SFI Research Centre for Data Analytics, University College Dublin, Dublin, IE
• Oxford Web Institute, Oxford University, Oxford, UK
• Deliberative Democracy Lab, Stanford University, Stanford, CA, USA
• Tepper School of Business, Carnegie Mellon University, Pittsburgh, PA, USA