The Future of AI Lies in Technology Collaboration
RISE researcher Anders Holst argues that the most significant AI advances will come not from single technologies but from combining multiple AI capabilities to solve complex real-world challenges.
Anders Holst, Researcher at RISE, believes that the most significant developments in AI will emerge not from any single technology, but from the combination of several different AI technologies working together to solve complex challenges.
The Core Insight
“Intelligence is not a single ability, but is made up of multiple abilities,” Holst explains. Just as the human brain is divided into centers for language, spatial thinking, abstract reasoning, motor skills, and social skills, effective AI systems need to combine multiple specialized capabilities.
The AI Technology Stack
Modern AI encompasses many distinct technologies:
- Machine learning and data analysis
- Computer vision
- Natural language processing (text and speech)
- Robotics
- Planning and optimization
- Knowledge representation
- Automated reasoning
While each can achieve impressive results in controlled environments - like 98% accuracy in image recognition - real-world applications demand integration.
The Integration Challenge
The key difficulty involves developing what Holst calls the “AI glue” - mechanisms that enable different technologies to communicate effectively within complex systems. This integration challenge is where much of the cutting-edge research is focused.
Practical Applications
Manufacturing environments represent a primary application area where combined AI systems are making a difference. Holst predicts that AI will make the greatest impact at the system level in factories, where planning, production, and maintenance can be optimized dynamically based on analyses from multiple variables rather than predetermined schedules.
RISE’s Track Record
RISE has pioneered multi-technology AI approaches in several projects:
- SADV (2010-2016): Combined machine learning with rule-based systems for maritime anomaly detection
- FUSE (2014-2016): Integrated statistical learning with planning for maintenance optimization
- MARE: Merged data fusion with machine learning for network error handling
Looking Forward
The future belongs to AI systems that can seamlessly combine perception, reasoning, and action - much like biological intelligence does naturally.
