Trustworthy AI: Access to Justice Interdisciplinary Research

Section 1Q1: Define the Problem

The Challenge We're Addressing

Digital Rights and Data Extraction

the information is often cataloged as work product and no easy way to get the data. Need to investigate current issues on this]. We have made this agreement that if they give the tool for free, they can get all of the data connected to the product, adjacent to the product, and inferred from usage of the product. Because of IoT, this data may be sold to other companies and other usages outside of the agreed terms of usage of the data, which are complicated to understand.

  • There needs to be transparency of human language (1 page or less) that allows an individual to decide if the tool’s utility is worthy giving up a lot of data about oneself. Records management. Europe has an opt-in model. US has an opt-in model. How long do these companies hold onto the data?

Algorithmic Fairness and Public Trust in Adjudication

Some early research shows that women and people of color prefer AI adjudication over humans, believing it will be fairer.


What kind of justice should AI work in changing/impacting?

The construction of AI environment dedicated to access to justice in a way that can analyze non-structured data in order to translate it into metrics of justice at different dimensions (micro, meso, macro)? Two tasks: complete the state of the art from the point of view of justice and from point of view of artificial intelligence and all together find the possible agents and then tailor proposal to what the program managers want.

Tribal sovereignty/rights with AI data collection/distribution and data literacy.

 

Is using the law to address AI regulation and guardrails, is that a pathway to reducing bias. Do we have lawmakers who are not only in a position to create guardrails, but they also purchase AI. Deciders who are participating in creating AI standards, are also consumers. In an environment where they are worried about their own liability and still making decisions. Dealing with a clash of rights.

How do we show that peace is profitable?

Section 2Q2: What is the Research?

Our Research Approach

Research Focus
user knowledge of data collected; user appreciation of data collected usage; terms of agreements, law and regulations related to data collection (CCPA 2018)], [domains]. [Research methods: literature review, quantitative data set collection of company data collection (Electronic Frontiers Foundation (EFF)), [frameworks], [tools: FOIA, data set access, programming for data mining], [strategic approach: data mining for patterns]
Key Research Questions:
  • What is the minimum information that a user will need to comprehend user agreements when it comes to technology tools along with AI (mobile apps, etc.)?
  • What happens to the data when it is collected?
  • Who do they share the data with? Do we have the right to be forgotten?
  • What is the minimum companies need to operate (though they collect more than what they need
Methodology & Approach
Justice, artificial intelligence, sociology, medicine, business, economics. Research methods: literature reviews of state of the art of foundational models to do the transduction, study sources of data (through a social center)
Section 3Q3: What Value Will the Research Provide?

Anticipated Impact & Benefits

This can give companies guidelines. Can give users agency when deciding how to engage with a company. Can help legislators to create quality policy.
[Description of specific value/impact]
from pure AI research, we would be working in the usage of foundational models to go beyond the pure prompting. Train AI models to extract the info that we want
[Description of specific value/impact]
Benefit Type 3
[Description of specific value/impact]
Section 4Q4 & Q5: Stakeholders & Audience

Who's Involved & Who Should Care

Key Stakeholders
Companies, New Mexico residents, legislators
Courts, legal system
Communities, city of Albuquerque (if they are interested in applying access to justice concepts to intervention plans)
Universities, governmental entities (city, state, county)
Target Audience / Beneficiaries
New Mexico residents, NM legislators, K-12
Section 6Q5: Identify Possible Funding

Potential Funding Sources

LFC
LFC Research Arm
EHH
[Grant/Type]
NIH
[Grant/Type]
NSF
[Grant/Type]
Section 6

Our Interdisciplinary Research Sub-Group

Sonia Gipson-Rankin
Co-Lead/Prof. School of Law
Manel Martinez-Ramon
Co-Lead/Prof. Electrical & Computer Engineering
Todd Quinn
Member/Assoc. Prof. CULLS
Kathy Powers
[Role/Title]
Trilce Estrada
[Role/Title]