Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Good Intentions, Risky Inventions: A Method for Assessing the Risks and Benefits of AI in Mobile and Wearable Uses

Marios Constantinides, Edyta Bogucka, Sanja Scepanovic, Daniele Quercia

TL;DR

This study introduces a semi-automatic pipeline that uses LLMs to enumerate AI uses in mobile and wearable devices, classify each use’s risk under the EU AI Act, and assess benefits against the UN SDGs. The approach combines three prompts to generate uses, evaluate risks, and determine benefits, with expert validation showing >85% accuracy and crowdsourcing revealing nuanced judgments across risk and SDG mapping. Key findings show many low-risk uses favor environmental sustainability and logistics, while several high-risk uses offer meaningful benefits in well-being, safety, and equality but raise concerns about sensitive data and automated decisions. To translate these insights into practice, the authors present a Risk Assessment Checklist tailored for the Mobile HCI community, aiming to balance risks and benefits without discarding impactful technologies.

Abstract

Integrating Artificial Intelligence (AI) into mobile and wearables offers numerous benefits at individual, societal, and environmental levels. Yet, it also spotlights concerns over emerging risks. Traditional assessments of risks and benefits have been sporadic, and often require costly expert analysis. We developed a semi-automatic method that leverages Large Language Models (LLMs) to identify AI uses in mobile and wearables, classify their risks based on the EU AI Act, and determine their benefits that align with globally recognized long-term sustainable development goals; a manual validation of our method by two experts in mobile and wearable technologies, a legal and compliance expert, and a cohort of nine individuals with legal backgrounds who were recruited from Prolific, confirmed its accuracy to be over 85\%. We uncovered that specific applications of mobile computing hold significant potential in improving well-being, safety, and social equality. However, these promising uses are linked to risks involving sensitive data, vulnerable groups, and automated decision-making. To avoid rejecting these risky yet impactful mobile and wearable uses, we propose a risk assessment checklist for the Mobile HCI community.

Good Intentions, Risky Inventions: A Method for Assessing the Risks and Benefits of AI in Mobile and Wearable Uses

TL;DR

This study introduces a semi-automatic pipeline that uses LLMs to enumerate AI uses in mobile and wearable devices, classify each use’s risk under the EU AI Act, and assess benefits against the UN SDGs. The approach combines three prompts to generate uses, evaluate risks, and determine benefits, with expert validation showing >85% accuracy and crowdsourcing revealing nuanced judgments across risk and SDG mapping. Key findings show many low-risk uses favor environmental sustainability and logistics, while several high-risk uses offer meaningful benefits in well-being, safety, and equality but raise concerns about sensitive data and automated decisions. To translate these insights into practice, the authors present a Risk Assessment Checklist tailored for the Mobile HCI community, aiming to balance risks and benefits without discarding impactful technologies.

Abstract

Integrating Artificial Intelligence (AI) into mobile and wearables offers numerous benefits at individual, societal, and environmental levels. Yet, it also spotlights concerns over emerging risks. Traditional assessments of risks and benefits have been sporadic, and often require costly expert analysis. We developed a semi-automatic method that leverages Large Language Models (LLMs) to identify AI uses in mobile and wearables, classify their risks based on the EU AI Act, and determine their benefits that align with globally recognized long-term sustainable development goals; a manual validation of our method by two experts in mobile and wearable technologies, a legal and compliance expert, and a cohort of nine individuals with legal backgrounds who were recruited from Prolific, confirmed its accuracy to be over 85\%. We uncovered that specific applications of mobile computing hold significant potential in improving well-being, safety, and social equality. However, these promising uses are linked to risks involving sensitive data, vulnerable groups, and automated decision-making. To avoid rejecting these risky yet impactful mobile and wearable uses, we propose a risk assessment checklist for the Mobile HCI community.
Paper Structure (21 sections, 3 figures, 1 table)

This paper contains 21 sections, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work and Background
  3. AI Regulations and Impact Assessments
  4. Risk and Benefit Assessments of AI for Mobile and Wearable Uses
  5. LLMs and Prompt Engineering
  6. Author Positionality Statement
  7. Methods
  8. Prompt #1: Generating Uses
  9. Prompt #2: Classifying Use Risks
  10. Prompt #3: Determining Use Benefits
  11. Crowdsourcing Study
  12. Results
  13. Uses Promoting SDGs and Being Low Risk
  14. Uses Promoting SDGs yet Being High Risk
  15. Understanding the Reasons Why a Use Is Low Risk, High Risk, and Beneficial
  16. ...and 6 more sections

Figures (3)

  • Figure 1: Crowdsourcing study survey. Participants were given instructions about the study (1) and were provided the definitions of risk classifications as per the EU AI Act (2) and the definitions of the 17 Sustainable Development Goals (3). Then participants were presented with 46 mobile HCI uses (4,5) and asked to assess how probable is the use (Q1), whether they agree with the LLM-generated risk classification (Q2) and its justification (Q3), explain their reasoning about the classification and justification (Q4), and select the SDGs that they believe the use supports. After annotating all uses, they were redirected to the Prolific confirmation page (6).
  • Figure 2: Percentages of uses categorized as low risk and high risk for each goal. The total number of uses per goal is in parentheses below the goal's name.
  • Figure 3: Risk Assessment Checklist for Mobile Computing. It helps to systematically consider a mobile or wearable system's: (A) use, (B) data, (C) risks, (D) mitigations and (D) compliance with legal and regulatory standards.