The Thirty-Ninth Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-27) is a premier venue for novel studies of AI deployments in real-world applications. IAAI-27 will feature technical papers, best practices, invited talks, and panel discussions that explore challenges, methodologies, and lessons learned from the practical use of AI. Submissions should focus on applied AI; purely theoretical work and algorithmic descriptions are more suited for AAAI-27.

IAAI-27 seeks novel contributions in the following areas:
(1) Deployed Applications showcasing the novel use of AI with measurable benefits, best practices, and lessons learned;
(2) Emerging Applications of AI on trajectory for full deployment;
(3) Tools and Methodologies that accelerate safer AI development and deployment and analysis of factors contributing to failures, incidents and their mitigation.

Tracks and Topics

1. Deployed Highly Innovative Applications of AI​

Papers submitted to this track must describe deployed applications that demonstrate measurable benefits through innovative use of AI technology. An application is considered deployed once it is in production and used by end-users, with meaningful data collected on its performance. Submissions may focus on either standalone applications or components of larger systems. Papers will be evaluated based on the quality of the problem description, AI approach, innovation in AI use, measurable improvements, deployment details, and lessons learned during development, deployment, and maintenance. In-depth algorithm descriptions are better suited for AAAI. Each accepted paper in this track will receive the IAAI ‘Deployed Application’ Certificate Award.

Page Limit: 8 pages, Format: AAAI style and formatting guidelines.

*No page limit for references and appendices.

2. Emerging Applications of AI

The Emerging Applications track focuses on novel applications of AI methods to real-world problems that are not yet fully deployed but include early deployment or pilot-stage results. Submissions must demonstrate a clear path toward full-scale deployment, addressing emerging engineering or sociotechnical challenges and their practical relevance. Papers will be evaluated on problem significance, innovation, AI methodology, technical quality, and clarity, with a clear path toward deployment. Submissions to this track have greater flexibility in how the papers are reviewed, but the quality threshold for acceptance is heightened with increasing distance from deployment. Papers without an articulated path to deployment may be rejected without review of the paper substance.

Page Limit: 6 pages, Format: AAAI style and formatting guidelines.
*No page limit for references and appendices.

3. Tools and Methodologies for Moving Faster and Safer

Within this track, we solicit papers describing deployed tools, practices, policies, and methods improving applied AI innovation and deployment of AI systems. Areas of interest include, but are not limited to:

Incident Analysis: Analysis of development and deployment processes and artifacts, focusing on real-world harm resulting from one or more AI incidents.
Incident Trend Analysis: Analyze trends in AI incidents to identify unresolved challenges in the deployment of AI systems.
Incident Best Practices: Detail best practices in use for producing and deploying systems to prevent or mitigate AI incidents.
Deployed Process Organization: Tools in use that help manage and assure the development, evaluation, or deployment of AI systems.
Deployed Data Hygiene and Data Quality Tools: Tools in use that are designed to address challenges in processing raw data in AI systems.
Deployed Meta-Optimization: Tools in use for enhancing AI systems through algorithm configurators, algorithm portfolios, and hyperparameter optimization.
Deployed novel computational models: Tools in use to exploit new computational hardware, such as neuromorphic processors, quantum computers, and other application-specific AI chips and systems.
Papers will be evaluated primarily by the following criteria: the extent to which the research presented advances real world application quality, development, deployment, maintenance costs, productivity, reliability, and/or scaling. Tools that are deployed will be ranked more highly in the review process. Analysis of failures must center on failures experienced during AI deployments (i.e., with systems directly impacting the real world). The evaluation will strongly factor in the degree to which the research is likely to enable real-world applied AI researchers to produce better AI solutions, including factors related to ease of use, adoption, documentation, and impact on the practices of the applications community.

Tools and methodologies still in the concept phase, or tested only in laboratory settings, are not suitable for this track. Responsive papers will demonstrate that use of the tool/methodology has led to measurable benefits.

Papers will be reviewed by a program committee of AI researchers, industry experts, human factors, safety, and ethics, with a particular focus on whether the paper efficiently informs the production and deployment of AI systems.

Page Limit: 6 pages, Format: AAAI style and formatting guidelines.
*No page limit for references and appendices.