Human-Computer Interaction for CSAI

by Albert Major

Building a Smarter Society

This lecture explores the intersection of Human-Computer Interaction (HCI), robotics, and ethics. It focuses on how technology can influence behavior (Persuasive Technology), how objects are connected (IoT), case studies of failure in robotics, and manipulative design (Dark Patterns).

1. Persuasive Technology

The goal of persuasive technology is to change user behavior or attitudes in a positive, non-deceptive way. It guides users to adopt better habits.

The Functional Triad (Key Concept)

Persuasive technology functions in three distinct ways. This is a critical concept for understanding how these systems operate:

• As a Tool:
  • Function: Increases the user's capability or efficiency.
  • Example: Smart Thermostats (e.g., Nest). It learns the user's schedule and temperature preferences to automatically save energy and optimize comfort. It makes the task of managing environment easier.
• As a Medium:
  • Function: Provides an experience, simulation, or environment to explore cause-and-effect relationships.
  • Example: Simulation games or VR applications. A first-aid training app allows users to practice skills in a virtual environment. "Serious gaming" in the military allows soldiers to train without physical risk.
• As a Social Actor:
  • Function: Creates a relationship with the user by leveraging social rules and emotional connections.
  • Example: Tamagotchi (digital pet) or Digital Coaches. Users feel a sense of responsibility to "feed" the digital pet or listen to the coach because the device mimics social behaviors, encouraging the user to return and engage.

2. Society 5.0 and The "Moonshot" Project

The lecture discussed a Japanese initiative (Moonshot Goal by 2050) aimed at creating a human-robot symbiotic society.

• Goal: To liberate humans from the limitations of:
  • Body: Using cybernetics or robotics to overcome physical disabilities.
  • Brain: Enhancing cognitive capabilities.
  • Space and Time: Allowing remote presence.
• Avatar Technology: A key component is the use of avatars that allow people to socialize, work, or travel remotely, effectively being in two places at once.
• Assistive Robotics: Concepts like "Nimbus" (wearable robotics/exoskeletons) support daily mobility, helping the aging population or those with injuries.

3. Internet of Things (IoT)

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

The IoT Formula

For a device to be considered IoT, it generally needs:

• A Thing: The physical object (e.g., a trash bin).
• Sensors: To collect data (e.g., detecting if the bin is full or detecting faces).
• Processor/Software: A microcontroller to process the sensory data locally.
• Network: A connection (Local or Internet) to transmit decisions or data.

Core Components of IoT Systems

• Low-Power Embedded Systems: Devices often run on batteries for years (e.g., Bluetooth sensors), requiring highly efficient power management.
• Cloud Computing: Used for processing heavy data loads. A single camera running at 60fps generates too much data for a small chip to handle; data is sent to the cloud for analysis.
• Network Connection (IP): Every connected device needs an IP address to be reachable. If you can "ping" the device, it is connected.

Applications

• Ambient Intelligence: The environment itself helps the user. For example, a hospital room with sensors that detect if a patient falls and automatically alerts nurses.
• Smart Warehouses: Used by companies like Ferrari or Mazda to track inventory and security using heat and motion sensors.

4. Case Studies: Failure in Social Robotics

Despite high budgets and top talent, many social robots fail because they do not solve a real problem better than existing tools (like tablets).

• Jibo:
  • Context: Raised over $63 million, backed by MIT experts, and featured on Time Magazine.
  • The Pitch: "The first family robot" that fits in like a family member.
  • Why it Failed: It didn't offer utility beyond what a smartphone or tablet could do. The novelty of a "social" device wasn't enough to sustain it as a product.
• Amazon Astro:
  • Context: Describe as "Alexa on wheels."
  • Critique: Often viewed as an expensive "cup holder." It attempts to integrate into the family but often fails to justify its high price point ($1,000+) with actual usefulness.

5. Dark Patterns

Dark patterns are the opposite of persuasive technology. They are interface designs crafted to trick or manipulate users into doing things they might not want to do.

Three Common Types:

• Nagging:
  • Definition: Persistent interruption that disrupts the user's workflow without a clear way to permanently say "no."
  • Example: Instagram notifications asking to be turned on. The options might be "OK" and "Not Now." There is no "Never" option, so the user is eventually worn down into accepting.
• Obstruction:
  • Definition: Putting barriers in the way of a user's intent, specifically to prevent them from cancelling a service or changing a setting.
  • Example: Hiding the "Limit Ad Tracking" option deep within multiple sub-menus (Settings -> General -> About -> Advertising) so users cannot find it easily.
• Interface Interference:
  • Definition: Visually manipulating the interface to privilege one action over another.
  • Example: An installation wizard where the "Express Install" (which might install junk software) is a big, highlighted button, while "Custom Install" is small or greyed out, making users think it's the wrong choice.

6. Ethical Dilemmas in Tech

The lecture presented several scenarios to highlight the difficulty of making ethical decisions in the workplace:

• The Overselling Boss: If a boss promises features to a client that do not exist, the engineer faces a conflict between honesty and job security.
• The Product Launch: If usability testing fails but marketing insists the product is ready, the designer must decide whether to alert the CEO (risking their reputation) or stay silent (risking product failure).
• Data Usage: A scenario where a researcher is offered a job using Facebook data collected without explicit user consent (similar to Cambridge Analytica). The dilemma is between career advancement and data privacy ethics.
• Inclusion vs. Exclusion: When building an online community (e.g., a book club), designers must decide if "inclusivity" means allowing everyone, or if excluding certain extreme behaviors is necessary to protect the safety and vibe of the core community.