Discovering Proxy Systems to Test Critical AI Systems: A Metadata-Driven Software Similarity Approach

Mark Kennedy; Joanna F. DeFranco; Phillip A. Laplante

doi:10.1109/MC.2026.3653167

Discovering Proxy Systems to Test Critical AI Systems: A Metadata-Driven Software Similarity Approach

Mark Kennedy
, Joanna F. DeFranco
, Phillip A. Laplante

Engineering Division (Great Valley)

Research output: Contribution to specialist publication › Article

Abstract

This article describes a practical approach for finding “software twins,” noncritical proxy systems that mirror critical artificial intelligence-enabled systems and can be safely abused during testing, even when proprietary source code cannot be accessed for direct comparison.

Original language	English (US)
Pages	142-146
Number of pages	5
Volume	59
No	3
Specialist publication	Computer
DOIs	https://doi.org/10.1109/MC.2026.3653167
State	Published - 2026

All Science Journal Classification (ASJC) codes

General Computer Science

Access to Document

10.1109/MC.2026.3653167

Cite this

@misc{854b1cb994e04cd2b2fc4980ea5f4ff2,

title = "Discovering Proxy Systems to Test Critical AI Systems: A Metadata-Driven Software Similarity Approach",

abstract = "This article describes a practical approach for finding “software twins,” noncritical proxy systems that mirror critical artificial intelligence-enabled systems and can be safely abused during testing, even when proprietary source code cannot be accessed for direct comparison.",

author = "Mark Kennedy and DeFranco, \{Joanna F.\} and Laplante, \{Phillip A.\}",

note = "Publisher Copyright: {\textcopyright} 1970-2012 IEEE.",

year = "2026",

doi = "10.1109/MC.2026.3653167",

language = "English (US)",

volume = "59",

pages = "142--146",

journal = "Computer",

issn = "0018-9162",

publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Discovering Proxy Systems to Test Critical AI Systems

T2 - A Metadata-Driven Software Similarity Approach

AU - Kennedy, Mark

AU - DeFranco, Joanna F.

AU - Laplante, Phillip A.

PY - 2026

Y1 - 2026

N2 - This article describes a practical approach for finding “software twins,” noncritical proxy systems that mirror critical artificial intelligence-enabled systems and can be safely abused during testing, even when proprietary source code cannot be accessed for direct comparison.

AB - This article describes a practical approach for finding “software twins,” noncritical proxy systems that mirror critical artificial intelligence-enabled systems and can be safely abused during testing, even when proprietary source code cannot be accessed for direct comparison.

UR - https://www.scopus.com/pages/publications/105030922207

UR - https://www.scopus.com/pages/publications/105030922207#tab=citedBy

U2 - 10.1109/MC.2026.3653167

DO - 10.1109/MC.2026.3653167

M3 - Article

AN - SCOPUS:105030922207

SN - 0018-9162

VL - 59

SP - 142

EP - 146

JO - Computer

JF - Computer

ER -

Discovering Proxy Systems to Test Critical AI Systems: A Metadata-Driven Software Similarity Approach

Abstract

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this