I Can’t Get No — Assistance. (Hey Hey Hey, That’s What I Say!)
Sprint 5: More Research, Insights, & Reframing
If you’ve been following along since the beginning of Sprint 1, you might have noticed our seemingly inexplicable devotion to pop music references for our articles. It’s safe to say that we enjoy music and hiding Easter eggs in our publications. (It’s ok if you haven’t noticed, but 5 points to whoever can reply back to us first with all of the songs in Sprints 1 through 5).
So, what does this have to do with our work on communication in space? Well, work is the key word here; sometimes, we can forget that we’re not just robots crushing out task after task in the most efficient way possible. We’re human, and so are the astronauts and members of mission control that we’re designing for. We realize that we need to consider not just usability and efficiency in our human-centered design thinking approach but also emotion, motivation, inspiration, creativity, and so many other human qualities. That said, we like to make sure our own work also exemplifies these values. This same energy and emotion guides our designs — Medium posts included ✌
Course Correction
Adjusting Our Research Methods 🧭
Tributes to one of the greatest rock bands ever aside (ahem.. The Stones), we are now cruising at terminal velocity towards the end of the spring semester in the MHCI program (see above for previous location). It’s hard to believe that we’ve only been on this project with NASA for 11 total weeks (feels more like 5 to 6 months).
Since our last post, we’ve adjusted course slightly to really explore some of our initial findings from speed dating. Instead of continuing the speed dating process with less and less unique findings from participants, we transitioned to conducting more simulated missions and facilitating discussions centered around understanding the high-level impacts of automation, artificial intelligence, and assistance in communication. This understanding is vital to the longevity of the near term communication features that we’ll be prototyping this summer.
With the end of our primary research phase rapidly approaching, we decided that investigating some of our far-reaching ideas (like intelligent assistants designed to aid crew communication) would provide much more insight and guide our future path, designs, and prototypes in a more fruitful way.
More Simulated Missions🌋
We prefer gathering live action data, so naturally we utilized our simulated missions for diving into our far-term ideas. We’ve come a long way with our protocol since the Lego days, and we’ve since moved on to a scenario that most closely resembles what an EVA — extravehicular activity — might look like. We’re recruiting local Pittsburgh participants to play the typical roles of IVs, EVs, and mission control. The IV (intravehicular) operator is inside the “habitat” coordinating tasks with the EV operators completing tasks out in the field, while mission control is observing and providing the necessary feedback from a distance and with a delay.
We’ve evolved our simulated mission protocol to quickly test our hypotheses in the name of rapid iteration and feedback. In addition to the usual Wizard of Oz-ing that we’re doing to simulate message delay between mission control and crew, we added another wizard element — the IV assistant. Casting a member of Team Chronos as the assistant, we’re studying how, when, and why users may or may not decide to ask for assistance in the context of the simulated mission. Knowing our results are only distantly representative of our users and a real EVA, we are using these missions not only to glean general insights about the potential use cases and constraints surrounding automation and AI in communication but also to refine our protocol for usability testing in the summer. We imagine that we can continue to apply our simulated missions for testing our various prototypes with more selective users this summer!
AI Discussions with Other MHCI Capstone Teams 🤝
Artificial intelligence has really burst onto the scene in recent years, and with that, comes plenty of misuses and misunderstandings. Sometimes, it gets that magic buzzword / black box treatment, and that’s exactly what we want to avoid when designing for assistance in communication.
Luckily, we’re surrounded by 11 other capstone teams of the best and brightest human-computer interaction students out there, and many of them are also grappling with the impacts and nuances of AI in the design space. In order to quickly consume and share ideas, we met with 3 different teams to discuss use cases and findings about artificial intelligence, machine learning, and conversational agents. Some of the discussion was as expected: AI/ML isn’t the solution for everything. However, we also learned some valuable lessons in that context, predictability, expectations, and user control are key factors to success of your designs. We only scratched the surface of the whole “trust in AI” problem in these discussions, but the real advantage is that we now have a much better vision for crafting our prototypes and ideating around the contexts that our well informed and highly technical users (👩🚀) might use them in.
Initial Findings on AI in Communication 🤖
In a short two weeks, we’ve gathered a lot of information on AI and how it might be used in the context of delayed communication in future deep-space missions. From our simulated missions, we found that crew members requested assistance in the following general areas:
- Time Management
- Progress Tracking
- Baseline Judgment and Recommendations
A key point to note in regard to this is that crew members weren’t looking for tasks to be done for them. Instead, they were looking for the type of guidance that would allow them to be strategic and make informed decisions to accomplish the mission. In addition, our cohort discussions revealed some other general requirements in regard to AI and automation. If you’re going to implement a machine to assist humans, it has to be:
- Perceivable — You want to be aware of what the machine can do and what it’s currently doing. In typical interaction design terms, your product needs to provide feedforward information of what it’s capable of doing and also its current state.
- Predictable — You want to know what the machine is going to do before it does it. Familiarity through feedback loops and a “conversation” between user and machine can help build a shared understanding. (Shout out to our Interaction Design Professor for inspiration here — Paul Pangaro)
- Controllable — You want to have the final say in what the system is doing and when. Being able to opt-in and opt-out of its services is key to not feeling undermined by it.
All of these concepts must come together in order to create a successful system that is trusted by its users, but the context in which the system is implemented, the tasks it’s designed for, and the characteristics of its users are also essential. The goals and the benefiting stakeholders must be clearly defined in order to know where the assistant fits in the bigger picture.
Our aim is to take these initial findings and shape our future plans for prototyping in the summer. Regardless of the current state and the solutions that will likely play a role in Playbook’s improvement in the near term, there is overwhelming evidence in our research that automation, artificial intelligence, and machine-based assistance must be involved in the future state where the weight of the mission success is on the shoulders of only a few crew members. There is no doubt that these small teams will need assistance in all aspects of the deep space mission — communication included.
Course Correction 2.0 (aka Next Steps)
Research, synthesize, pivot. Research, synthesize, pivot. 🔁
That’s the name of the game, that is the user-centered research and design thinking process. The growing success of our simulated missions means it’s time to start something new. Funny, right?
To wrap up the research phase of our capstone (beginner tip: research never really ends), we’ll be running one more simulated mission with some researchers and designers at NASA AMES, doing a full-blown roundup on our AI findings, and then spending the majority of Sprint 6 doing competitive analysis while also preparing for our project visioning this summer. Look out for our Sprint 6 post for more details!
Chronos Communication
We can’t forget about our Chronos Communications on the really cool, really fun, really awesome things we did and found during Sprint 5. We’ll leave you with a glam shot of our 3D printed rocket from Thingiverse and a video of Simone Giertz being locked in a bathroom for 48 hours. We will NOT be asking our participants to do this, but that’s some next-level dedication to empathy building with future Mars astronauts!
Signing off for now,
Chronos
Chronos Acronym Dictionary
AI: Artificial Intelligence
ARC: Ames Research Center
BASALT: One of NASA’s analog mission project to design and develop elements of future missions that could send humans to conduct science and exploration on Mars
CAPCOM: Capsule Communicator
CCTV: Closed Circuit Television
CDMS: Command & Data Management Systems Officer
CDR: Commander
CDS: Central Data System
DCS: Display & Control Monitor
ESA: European Space Agency
EVA: Extra-Vehicular Activity
F/C: Flight Controller
FD: Flight Director
HERA: Human Exploration Research Analog
IMF: In Flight Maintenance
INCO: Instrumentation & Communications Officer
IVA: Intra-Vehicular Activity
JSC: Lyndon B. Johnson Space Center
KSC: John F. Kennedy Space Center
MCC: Mission Control Center
MD: Mission Director
MS: Mission Specialist
MSCI: Mission Scientist
NEEMO: NASA Extreme Environment Mission Operation
SA: Situational Awareness
SME: Subject Matter Experts
Opinions expressed are solely our own and do not represent the views or opinions of The National Aeronautics and Space Administration (NASA)
