About

AdviceIT

A research simulation about explaining AI investment advice, built by a team of three. It exists to find out which kinds of explanation help people follow good advice and catch bad advice. Nothing on this site is real financial advice.

What AdviceIT is

The site has a working AI advisor that recommends an investment mix for a made-up person. It also has several ways of explaining that recommendation. People who take part read six short cases, see the advisor's recommendation with one kind of explanation, and say what they would do. Half of the recommendations are deliberately wrong, and the study measures whether the explanation helped people notice.

The advisor learned from 400 cases that were reviewed by a panel of four financial experts. Everything runs in your browser, nothing about you is collected, and the whole thing is open source.

The team

AdviceIT is a team research project from the Department of Information Systems, Institut Teknologi Sepuluh Nopember, Surabaya. It grew out of a systematic literature review the three of us wrote together on trust and algorithm aversion in the choice between human and AI financial advisors.

Raditya Pratama

Lead developer and research design

Built the website, the two advisors, the explanation methods and the study flow, and leads the research design.

Fausta Irsyad Ramadhan

AI ecosystem and core resources

Provides the infrastructure, compute and tooling the advisors and the site are built and served on.

Muhammad Wahyudi Wicaksono

Quality assurance and research validation

Tests the instrument against the research objectives and reviews the flow, the wording and the measures.

Roles as agreed within the team. Author order on any paper is decided per paper.

For researchers and reviewers

Design, data and methods

The study design, the training data, the model results, the references and the collected responses are each published on their own page.

Version 2.15.0. To cite the instrument, use the citation file in the repository.

Take part in the study

About fifteen minutes, anonymous, and no real money involved.

Take part in the study