Interdisciplinary collaboration in higher education is more crucial than ever. By bringing together diverse viewpoints, not only across academic disciplines but from business and government too, we can drive innovation and tackle the complex challenges we’re facing. But fostering effective communication and knowledge sharing across disparate groups can often be difficult.
Perhaps AI advancements can offer a solution. Emerging players, such as DeepSeek from China, are delivering high-quality AI services at remarkably low costs. Other open-source models from companies like Meta also allow users to download and customise technology to meet their needs. Using technology to facilitate collaboration across disciplines could be a game-changer.
1. Effective data sharing and integration
By connecting AI via agency software and model context protocol, institutions can now seamlessly share data, findings and insights across diverse fields in a stealth mode. Without showing the actual raw data, which can be sensitive owing to privacy (or similar reasons), AI facilitates the exchange of information by pre-aggregating, or summarising, the meaning of data from various sources, without breaching data laws or regulations.
2. Identifying commonalities and differences to enhance communication
Generative AI can analyse extensive datasets to uncover common themes and differences across interdisciplinary studies, revealing insights that might otherwise go unnoticed. This capability encourages innovative thinking and a deeper understanding of complex issues. By producing human-like text and summarising complex information, AI enhances communication and improves clarity. This ensures all parties are aligned and leads to more effective collaboration and decision-making.
3. Personalised tools for collaboration with innovative business models
An AI model can be seen as a container of research knowledge – a newly available tool that transforms intangible intellectual assets into tangible formats. Rather than licensing traditional intellectual property (IP), universities now have the opportunity to license AI models themselves as a means of generating income. These models can be trained on specific data, allowing teams to create customised tools that address their specific challenges. This tailored approach allows for more relevant insights and solutions, boosting productivity and creativity in collaborative efforts.
The future of AI: the role of small models
As larger models continue to evolve, smaller models are emerging as vital contributors to interdisciplinary collaboration. Here’s how:
- Privacy and security: Small models can be deployed locally, addressing privacy concerns and protecting sensitive data. This localised processing enhances security, making it easier for teams to safeguard proprietary information.
- Accuracy in specialised areas: Small models can be fine-tuned for specific domains, addressing common issues like hallucination in generative AI. By specialising, these models provide more reliable insights.
- Economical resource use: Training and operating small models is generally more cost-effective than larger counterparts, which require extensive resources. This cost-efficiency encourages broader collaboration, enabling a wider range of institutions to participate in data sharing and joint research.
Steps to achieve enhanced collaboration through AI
To implement AI solutions for interdisciplinary collaboration, consider the following steps:
- Develop a foundation model: Create or leverage a robust foundation model that will serve as the basis for deriving various small models tailored to different stakeholders’ needs.
- Build communication layers: Establish user-friendly interfaces that allow stakeholders to interact with small models. This layer should facilitate seamless communication and data exchange among users.
- Enable interconnectivity among models: Ensure that small models can connect and communicate with each other. Implementing technologies like blockchain and federated learning can enhance data security and collaborative learning.
- Implement continuous reinforcement learning: Establish a system for ongoing reinforcement learning, where models can learn from one another and adapt to the rapidly changing landscape of interdisciplinary collaboration. This iterative process will improve the accuracy and relevance of AI solutions over time.
As we look to the future, the evolution of AI – including advancements in agentic AI and the potential for artificial general intelligence – holds immense promise for enhancing interdisciplinary collaboration. By leveraging its capabilities, particularly through the use of small models, we can improve cooperation across teams and drive innovation together.
Raymond Chan is assistant director for entrepreneurship at Hong Kong Baptist University.