Bots are able to perform repetitive actions, mimic human interaction and understand the world through sensors. Most of the existing bots are designed to serve individual users rather than integrating them as part of a group and attend different petitions by understanding the context and keeping track of the group task flow. The motivation is to understand which drivers are important to guarantee effective crowd interaction with bots and provide guidelines to platform designers. I have studied how bots can be useful in human environments such as education, social good, workplaces, and crowd marketplaces.
2. Motivation
● Understand which drivers are important to guarantee effective crowd interaction
with bots and provide guidelines to platform designers.
● I have studied how bots can be useful in human environments such as
○ Education
○ Workplace
○ Workarea
○ Marketplaces
4. Bots in the Workplaces - TaskBot - Problem
● Effective task management is essential to successful team collaboration.
● Teams formulate, discuss, refine, assign, and track the collaborative tasks over
communication channels, yet they must leave these channels to update their
task-tracking tools, creating a source of friction and inefficiency.
5. Bots in the Workplaces - TaskBot - Solution
To address this problem, I explored how bots might be used to mediate task
management for individuals and teams. I implemented TaskBot a chatbot that helps
teams manage their tasks from within platforms they use to communicate. TaskBot
mediates the communication and keeps track of the task progress.
6. Bots in the Workplaces - TaskBot - Results
I deployed the bot to eight different teams of information workers. I derived seven
insights for the design of future bots for coordinating work, such as: Hierarchical
task-assignment, Supporting self-communication, Dealing with human ambiguity, ,
Handling multi-threaded conversations, among others.
8. Bots for Learning - MATT - Problem Statement
● Critique is important to improve creative work and help learners of design to grow.
The “gold standard” of critique involves in-person discussion.
● Scaling expert critique is difficult as experts are scarce, have limited time and
privacy concerns.
9. Bots for Learning - MATT - Solution
To enable at scale access to expert critique I designed MATT, a chatbot that
micro-guides experts to critique in short bursts of time with mediated communication to
address experts' time and privacy concerns.
10. Bots for Learning - MATT - Results
● I conducted a field experiment comparing MATT to online forum, and to bot no
guidance.
● MATT's conversational micro-guidance facilitated at scale collaborations between
experts and learners.
11. Bots in the Workarea
ExperTwin:
An Alter Ego in Cyberspace for Knowledge Workers
SmartData 2018
12. Bots in the Workarea - ExpertTwin
● While organizing the entire Web as a semantic network is a long-term goal,
creation of a personal semantic network is an attainable task.
● I present ExperTwin, a personal and group botnet able to bring relevant
information to the workarea while performing a knowledge-based task.
13. Bots in the Workarea - ExpertTwin - Solution
ExperTwin brings relevant information to the workarea when needed and is shareable
with co-workers to create a collective knowledge base.
15. Crowd Marketplaces - META-GIG - Problem
● There are many platforms (e.g. Amazon Mechanical Turk, Uber, Upwork, Airbnb)
that provide access to different crowd markets.
● While it is common to use these sites as either a worker or requester, few of these
users have the power to create new markets -- meaning that existing markets may
not represent all users' needs and concerns.
16. Crowd Marketplaces
META-GIG - Solution
I present an evaluation of “META-GIG”, a platform that
enables users to create new crowd markets.
17. Crowd Marketplaces - META-GIG - Results
● The proposed ideas involved a range of occupations, social good components,
and advocated for workers' rights.
● We conclude by providing 8 design-implications for future crowd market creation
tools, such as: working conditions, local development catalyst, among others.
18. Next Steps
● META-GIG large deployment.
● Study how sharing economy requesters react to services offered by automated
processes.
19. Sequence
● Bots can empower people to learn and get critique from experts,
● Then people can have access to/create opportunities in crowd marketplaces,
● Then people can implement bots in their work area to be more efficient,
● Then people can work in teams and delegate task management to bots,
● Then people can delegate tasks to bots/people in crowd markets (next step)
20. Challenges
In order to generate insightful implications for design for the “Bots for Crowds” thesis
from the platform designer perspective, there are some challenges:
● I want to explore how bots can be implemented in marketplaces that allow
humans and bots, my challenge was to build first a platform that enables it and
then study it. How valid is that?
● I am unsure of what to measure in the large deployment of META-GIG platform.
● Unsure if the way in which I’m connecting my research is convincing.
● How to generalize the guidelines of each. I am studying bots not only chatbots.
● Should I discard/add some research to have a cohesive narrative?
● How can I better motivate my work in terms of Crowdsourcing and Artificial
Intelligence.