What are the implications for decision-making when we don’t know what we don’t know?
If there’s one thing we are all familiar with today, it’s uncertainty. At ANU, we have researchers who investigate how we manage uncertainty and decision-making in uncertain circumstances. Their goal is to advance Australian industries and determine how artificial intelligence will impact this.
We asked InSpace Mission Specialist Dr Zena Assaad, a Lecturer at 3Ai (the Autonomy, Agency and Assurance Institute) within the School of Cybernetics in the ANU College of Engineering and Computer Science, to tell us more about her work studying uncertainty, AI and how decision-making theories can help us understand how we manage our changing world.
Your studies focussed on decision-making under uncertainty. How does what you learned apply in the space sector and beyond today?
My research explored decision making under uncertainty, specifically for long-range air traffic flow management. Long-range flights can take upward of 10 hours. The information the pilots had when they took off has a high probability of changing along the course of the flight. Understanding the inherent uncertainty in flight paths and exploring methods of strategic decision making in those uncertain environments is key to enabling efficient and effective flight paths.
I proposed a decision-making theory for people faced with uncertain information who are also trying to balance multiple objectives. Balancing multiple objectives is complex and will ultimately lead to trade-offs in the final decision. In the end, complexity and trade-offs are magnified when the information people use for decision-making is imperfect and subject to change.
It is important when you consider the Australian civil space industry operates under uncertainty while attempting to manage many objectives. These objectives include national policy, supporting innovation and growth and leading international civil space engagement. Like many industries, the future is uncertain, and the information we have today has a high probability of changing or evolving over time. When you add in the new capabilities created by AI-enabled technologies, like design engineering assistants that use AI to support and reduce the time of initial mission design, this balance becomes increasingly more important.
Everyone today spends a lot of time thinking about how we pursue the best possible outcome under uncertain conditions with so many variables. When it comes to civil aviation, our future includes navigating the rapidly advancing landscape of the civil space industry and AI-enabled technologies more broadly. It becomes increasingly important to pragmatically assess the trade-offs of decisions made across different moments in time. As information evolves and becomes more accessible over time, how we interpret and act on that information will also change.
Your research explores emerging AI-enabled technologies and their potential impacts. Those impacts may be critical, but they’re not always obvious. How do you explain or show people the importance of your research?
Being able to explore the possibilities and potential impacts of emerging AI-enabled technologies is exciting and challenging. I enjoy being able to research, analyse and assess the possibilities of these technologies we hear about so often. For example, there is a lot of excitement and enthusiasm for drone delivery services. But what are the realities of having drones flying around suburban areas at low altitudes? What will it look like when we have a large number of drones operating in these conditions?
The reality is that, in a lot of cases, we are not yet at the point of seeing the full capability of emerging AI-enabled technologies. Attempting to replicate qualities related to human cognition, such as understanding, learning and problem solving, is a deeply complex and nuanced challenge.
The truth is today’s research, particularly around future social impacts, unintended consequences, safety considerations, etc., is exploratory research that is conducted against existing bodies of knowledge. This research is forward-looking and likely a reasonable indication of what’s to come. However, it can’t provide absolute certainty of any outcome.
The importance of forward-looking research is that it helps us make better decisions now and into the future. We can pursue research in anticipation of advancements and capabilities rather than in response to a problem or incident. Proactive research is critical in the pursuit of scaling technology capabilities and industry sectors at large.
What about space? Many people, understandably, have quite a narrow view of what space is and where it might take us–what does your research help tell us about industries like space?
Australia’s civil space industry is relatively new compared to other countries. Similar to other countries, we have a wide array of responsibilities beyond launch capabilities. These include regulatory systems, national safety issues and fulfilling national and international obligations. This wide net cast by the civil space industry requires a diverse set of research, skills and expertise to fulfil.
One of the most exciting areas of our space industry is the use of autonomous augmented reality to help train crew members on spacecraft repair and maintenance.
My research into the regulation and safety of autonomous systems creates a good opportunity for further investigation through a civil space lens. AI seems like a great opportunity to meet demands around scale, communication and reduced operating costs.
How much should we encourage increased integration of autonomous systems within the civil space sector? The integration of autonomous systems with human roles in space flight will pose challenges that can inhibit adoption and acceptance. I’m excited by the opportunity to further this research and its potential applications and implications associated with the civil space sector.
You can read more about Zena’s work at www.zenaassaad.com.