Human-Computer Interaction for CSAI

by Albert Major

Building a Smarter Society, IoT, and Ethical Design

This lecture focuses on the intersection of Human-Computer Interaction (HCI) and society, covering Persuasive Technology, the Internet of Things (IoT), case studies of failed robots, and deceptive design practices (Dark Patterns).

1. Persuasive Technology

Persuasive technology is designed to change user attitudes or behaviors through persuasion and social influence, rather than coercion or deception. The goal is to guide users toward positive behavior changes (e.g., saving energy, learning skills).

The Functional Triad: There are three ways technology functions to persuade users. This is a core concept for the exam.

• As a Tool (Augmentation):
  • Function: Increases capability or makes a target behavior easier to perform.
  • Mechanism: It reduces barriers (time, effort, cost) or automates decisions.
  • Example: A Smart Thermostat (like Nest). Instead of manually adjusting the heat, the device learns your habits and automatically adjusts the temperature to optimize energy consumption. It changes your behavior (energy saving) by automating the process.
• As a Medium (Experience):
  • Function: Provides an experience allows users to explore cause-and-effect relationships.
  • Mechanism: Uses simulation or visualization to teach new skills or show consequences.
  • Example: Virtual Reality First Aid training or military simulations (e.g., "The Guy"). By simulating a scenario, the user learns skills and behaviors in a safe environment that translate to the real world.
• As a Social Actor (Relationship):
  • Function: Creates a relationship with the user.
  • Mechanism: Uses social cues (praise, rewards, punishments) to invoke social responses from humans.
  • Example: Tamagotchi or a Digital Coach. A Tamagotchi requires care; if you neglect it, it dies. The user feels a social responsibility to the device, promoting the behavior of "caretaking." A digital coach uses social pressure (encouragement/disappointment) to motivate exercise.

2. Robotics and Society: The Japanese Moonshot Project

The lecture introduced Japan’s "Moonshot Goal 1" (targeted for 2050) as a response to an aging population and labor shortage.

• Goal: To liberate humans from the limitations of Body, Brain, Space, and Time.
• The Concept: Instead of forcing young people to work solely in elderly care, society uses technology to allow anyone to participate remotely.
• Key Technologies:
  • Cybernetic Avatars: Robots that allow a person to be telepresent. A person in Osaka could control a robot in Berlin to socialize or work.
  • Exoskeletons: Wearable robotics (referenced as "Nimbus" or similar to anime tech) that support mobility for the elderly or injured, functioning like an external muscular system.
  • Assistive Robotics: Creating a "symbiotic society" where AI and robots support daily living.

3. Internet of Things (IoT)

IoT refers to physical objects ("things") embedded with sensors, processing ability, and other technologies that connect and exchange data with other devices and systems over a network.

Definition & Logic:

• Equation: Thing + Sensor + Processor + Network = IoT.
• The "Internet" Misnomer: The device does not necessarily need the global Internet. It only requires a Network connection (could be local LAN, Bluetooth, etc.) and an IP Address to be addressable.

Core Components of IoT:

1. Sensors: The inputs (e.g., temperature, distance, cameras). Cameras generate massive amounts of data (e.g., 60fps x 1024px), requiring significant bandwidth.
2. Low-Power Embedded Systems: The hardware must be energy efficient (e.g., a watch battery lasting 10 years).
3. Network Connection: Every device serves as a node with an IP address (demonstrated via the ping command in the lecture) to communicate status.
4. Cloud Computing & Big Data: Due to the high volume of sensor data, processing often happens in the cloud rather than locally on the device.

Software Architecture Layers:

• Sensing Layer: Collects raw data (e.g., "motion detected").
• Business Logic Layer: The code that decides what to do with the data (e.g., "If motion is detected AND time is > 2:00 AM -> Alert Security").
• Application Layer: The interface the human sees (Frontend dashboard, CLI, or VR interface).

Ambient Intelligence: A sub-field of IoT where the environment itself acts as the "robot." For example, a hospital room equipped with sensors to detect if an elderly patient falls, automatically alerting caregivers without the patient wearing a device.

4. Case Studies in Failure: HRI (Human-Robot Interaction)

The lecture analyzed high-profile failures to demonstrate that money and talent do not guarantee success if the social context is misunderstood.

• Jibo (The "First Family Robot"):
  • Background: Created by Cynthia Breazeal (MIT), raised $63M+. Featured on Time Magazine.
  • The Promise: "Part of the family," distinct from a tablet, warm, and friendly.
  • The Failure: It over-promised and under-delivered. It was functionally just a stationary tablet/smartphone. Users realized it didn't solve a unique problem.
• Amazon Astro:
  • The Product: An "Alexa on wheels."
  • The Critique: Described by reviewers as a "$1,000 cup holder."
  • The Lesson: Companies try to force robots into the role of "family member," but users often find this creepy or useless. Understanding the HRI Heuristics (user expectations vs. reality) is critical.

5. Dark Patterns

Dark patterns are the ethical opposite of Persuasive Technology. They are UI/UX design choices specifically crafted to deceive users or force them into actions they did not intend to take.

Three Key Types (Exam Topic):

• 1. Nagging:
  • Definition: Persistent interruption without a clear path to permanently decline.
  • Example: Instagram notifications asking to be turned on. The options are usually "OK" and "Not Now." There is no "No/Never" option. The system nags you until you eventually give up and accept.
• 2. Obstruction:
  • Definition: Making a process unnecessarily difficult to discourage the user from doing it.
  • Example: "Limit Ad Tracking" on older iOS versions. The option was buried deep inside Settings -> General -> About -> Advertising, making it hard to find and disable.
• 3. Interference:
  • Definition: Visually manipulating the interface to privilege one option over another, often confusing the user.
  • Example: Software installers where "Express Install" (which installs unwanted bloatware) is a giant green button, while "Custom Install" (the clean version) is a tiny, grey text link.

6. Ethical Dilemmas

The lecture posed several scenarios to consider regarding professional responsibility:

• The Product Launch: Marketing says a product is ready; you (the designer) know it isn't usable. Do you tell the CEO? (Dilemma: Duty to user vs. job security).
• The Facebook/Cornell Study: A real researcher (Hancock) used Facebook data to manipulate user emotions (emotional contagion) without explicit consent. Would you accept a PhD position using such data? (Dilemma: Academic advancement vs. research ethics).