Human-AI Hybrid System for Early Career Exploration
In early career exploration under more uncertainties, conversational AI robot is a good assistant service.
*Selected to give the presentation at 2021 AMI&E Conference.
HAI research / HCI / UXR / UXD / User test / Woz / Conference
Issues on career exploration
Students often experience uncertainty anxiety caused by diversity, rapid change, and ambiguity, let them difficult to clarify goals, and easier to Burnout.
#1 Discover by Research
Grasping the essence of our subject through indirect investigation: pinpoint 5 crucial related studies out of 25 articles.
Explore with issues in mind: Field observation by AEIOU framework and shadowing.
Having deeper understand of what and how TA thoughts:
Semi-structure interview with 5 students.
#2 Define by Synthesis
Identify the main theme of TAs' needs, pain points:
5+5+3 hours Thematic analysis by affinity diagram.
Having clear view of TA's experiences by visualizing: Build User Journey Map and scenarios.
Establish the How Might We:
Give report to class, discuss with professor and get feedbacks to identify the right problem.
HMW1
make it easier for students to clarify their own thoughts about their career path by exploring future careers?
HMW2
make job information more transparent to help students use them widely and wisely?
HMW3
make job application preparation more efficiently?
#3 Develop by Ideation
Individual idea search, collection and brainstorming.
Build ideation design card deck together.
Collaborate ideation by ideation design card.
Identify the key solution by weighting matrix.
#4 Deliver by Implementation
Hypothetical solution
Human-AI chatbot assists.
automates and assists note-taking.
automatically generates relevant visual images to assist thinking (e.g. Mood-board).
Solution process
Assisting task & goal splitting.
Automated online resources recording and packaging.
Visual assistance while achieving goals.
#5 Feedback & Iteration
Do user testing to gain the feedbacks from TA:
Conduct 5 user test by giving the scenario to participant, and ask them to interact with our prototype built by Notion and LINE, with tester interacting with and simulating in another room.
Participant are ask to do 3 task and maintain the "Think aloud" method:
Build the vision broad of career goal.
Plan the task under his/her goal.
Read the weekly resources report and review his/her own progress.
After the test, interview with them to understand interesting points and by structure questions.
Also let participant fill out the SUS questionnaire to have quantitative result on how they feel about the service.
Finding
The needs of non-traditional students for the early career exploration.
Doner is more helpful for people who have ambiguous goals or are not good at organizing their own lives.
Conversational agent-Doner can be helpful to the first step of exploration.
Novelty: combine the Al agents with conversational apps into the progress of early career exploration
Research Discussion
AI assistance is particularly suitable for non-traditional education students and who don't set short & medium-term goals clearly.
AI assistance is more effective in early exploration and resource collection;
Innovation: combining AI agent and Conversational App(line) for early application Career exploration.
*Selected to give the paper presentation at The 5th international conference on Ambient Intelligence and Ergonomics in Asia.
Learning
Wizard of Oz is very suitable for user testing of AI-type products.
The matching between human-computer interaction and process under system module is crucial.
AI simulation of "工人智慧" has a high level of freedom and can properly test, feedback and make adjustments.
Special thanks
To teammates: Shih-Chu Chen, Ching-Yi Lai, Xin-Chang Li, Jie-Yu Huang
And prof. Yuan-Chi Tseng