We have long understood that humans learn by collecting and analyzing information. Machines also learn in this way, and because they do it far more efficiently than humans do, computer scientists have been teaching computers how to collect, process and find patterns in vast quantities of data. For the last 50 years, machines all around us have been learning, and societies across the world have embraced increasingly sophisticated technologies that influence how humans control their physical and virtual environments.
Machine learning will transform the world—more than it already has, and new technologies can have both positive and negative consequences. For instance, as self-driving vehicles and facial recognition systems become embedded into our everyday lives, how will the general public know if these emerging technologies meet ethical and safety requirements? By working with research and academic institutions, Underwriters Laboratories is helping to support the development and application of machine learning technologies, while also helping to ensure that the broader discipline of artificial intelligence is held to a standard of safety that makes the world’s digital future a safer one.
To investigate machine learning’s ongoing and potential impacts on human health and safety, Underwriters Laboratories and Northwestern University have launched the Machine Learning Impact Initiative, which aims to bring together a diverse network of researchers and practitioners to explore the following:
- Where should research be focused to better understand the risks and hazards related to machine learning?
- How can we better understand the inherent risks within the technology itself?
- How is machine learning applied to specific disciplines or areas of life?
- Where might the greatest impacts to individual human health and societal safety occur?
The Impact Initiative supports collaboration among individuals and groups who are building and researching the underlying algorithms and technical components of machine learning, those who are working in the frontiers of applying machine learning to various fields, and those who are studying its impact on humanity and society. “Underwriters Laboratories seeks to establish a community of inquiry,” said Terry Brady, President and CEO of Underwriters Laboratories. “With multiple voices represented in this dialogue, scientists can raise questions and help each another identify areas where further exploration is needed. Over time, the research collaborations will bring additional clarity to the broader field of artificial intelligence.”
“Northwestern, with its focus on cross-disciplinary work and impact, is well-suited to this task,” said Kristian Hammond, Professor of Computer Science at Northwestern University, “Working across our different schools, we are looking beyond what these technologies can do, so that we systematically consider what they mean to the human condition.” The Impact Initiative draws upon Northwestern’s leadership in artificial intelligence and seeks to build a diverse community of scientists, engineers and other specialists, drawing from a cross-section of institutions.
As its near-term goal, the collaboration aims to create a network of researchers who will study the issues related to the human and societal impacts of machine learning, followed by a roadmap that defines important focus areas. The group will explore whether the emerging technologies entrusted to learn on our behalf are doing so in a way that meets safety and ethical requirements. Over the long term, the teams will establish a research hub, one that will draw together professionals from all relevant disciplines to guide and support this work, and also to provide a central point for thought-leadership.
If you are affiliated with a research or academic institution and are interested in participating in this research, please click here.
- Underwriters Laboratories and Northwestern University are partnering to develop a methodology to assess the impact of machine learning applications on human health and safety in both theory and practice.
- If you are affiliated with a research or academic institution and are interested in participating in this research, please click here.
- Machine Learning uses large datasets to find patterns and features, and then make predictions based on them.
- In our increasingly data-rich world, the World Economic Forum predicts that 463 exabytes of data will be produced per day by 2025.