Deepfakes and international humanitarian law in armed conflict: a systematic analysis of documented cases and development of a legal compliance framework

Generative AI has created the capability to produce convincing video, audio and images of actual individuals and synthetic media of this sort has already been used as a tool of information warfare in actual armed conflicts. This poses a legal dilemma under international humanitarian law (IHL): international law provides for existing principles of distinction, perfidy, proportionality and precaution, which were formulated in an era when fabricated information was slow and costly to produce, and it remains uncertain how these principles would apply to information operations that are the product of AI, or whether there is a uniform approach to assessing documented cases. This gap is filled in this work by a doctrinal study of the issue in a qualitative manner, supplemented by a systematic comparative case analysis of eight documented examples of deepfake misinformation from the Russia-Ukraine conflict, the Israel-Gaza conflict and the Sudan conflict using information sources from independently verified fact-checking organizations, a state-funded disinformation-monitoring organization, and established international media. Ten legal and harm variables were coded for each case on a 4-point ordinal scale, and three attribution categories were coded for each case. The findings reveal that false attribution – making an attribution from a named leader that is not actually made – was concentrated in cases (three of eight cases were given a high score), proportionality concerns were widespread but not of a high level (six of eight cases received a moderate score on proportionality concerns), and attribution difficulty would be high in six of eight cases, with none receiving a public and court-ready attribution finding. In an article building on these results, it is proposed that a new framework be developed for assessing IHL compliance of the Deepfake called the Deepfake – IHL Compliance Assessment Framework (DICAF), which moves from content and purpose to the harm it causes and who it is attributed to before finally to the assessment of whether it meets or does not meet the requirements of IHL. This framework applies in detail to four of the eight cases. The article concludes that the existing principles of IHL do not lack applicability to this sample, but that they do not provide an operational way to categorize information operations generated by AI and that gaps in ICRC interpretive guidance for information with no evident attribution source and standards for verifying content are the most pressing gaps this sample presents to IHL

Authors

Publication Details

Journal
Interpersona An International Journal on Personal Relationships
Published
2026-11-13
DOI
https://doi.org/10.54899/iij.v20i4.1682
Primary Topic
International Law and Human Rights
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Deepfakes and international humanitarian law in armed conflict: a systematic analysis of documented cases and development of a legal compliance framework

Ahmed O.K Hosnia, Shihab Sulaiman Abdalla Osman, Randa Osman Ahmed Abd Alwahab, Hanan Abdel Moniem Gubara Elbashir
Interpersona An International Journal on Personal Relationships
International Law and Human Rights
article

Deepfakes and international humanitarian law in armed conflict: a systematic analysis of documented cases and development of a legal compliance framework

Ahmed O.K Hosnia, Shihab Sulaiman Abdalla Osman, Randa Osman Ahmed Abd Alwahab, Hanan Abdel Moniem Gubara Elbashir
article en

Abstract

Generative AI has created the capability to produce convincing video, audio and images of actual individuals and synthetic media of this sort has already been used as a tool of information warfare in actual armed conflicts. This poses a legal dilemma under international humanitarian law (IHL): international law provides for existing principles of distinction, perfidy, proportionality and precaution, which were formulated in an era when fabricated information was slow and costly to produce, and it remains uncertain how these principles would apply to information operations that are the product of AI, or whether there is a uniform approach to assessing documented cases. This gap is filled in this work by a doctrinal study of the issue in a qualitative manner, supplemented by a systematic comparative case analysis of eight documented examples of deepfake misinformation from the Russia-Ukraine conflict, the Israel-Gaza conflict and the Sudan conflict using information sources from independently verified fact-checking organizations, a state-funded disinformation-monitoring organization, and established international media. Ten legal and harm variables were coded for each case on a 4-point ordinal scale, and three attribution categories were coded for each case. The findings reveal that false attribution – making an attribution from a named leader that is not actually made – was concentrated in cases (three of eight cases were given a high score), proportionality concerns were widespread but not of a high level (six of eight cases received a moderate score on proportionality concerns), and attribution difficulty would be high in six of eight cases, with none receiving a public and court-ready attribution finding. In an article building on these results, it is proposed that a new framework be developed for assessing IHL compliance of the Deepfake called the Deepfake – IHL Compliance Assessment Framework (DICAF), which moves from content and purpose to the harm it causes and who it is attributed to before finally to the assessment of whether it meets or does not meet the requirements of IHL. This framework applies in detail to four of the eight cases. The article concludes that the existing principles of IHL do not lack applicability to this sample, but that they do not provide an operational way to categorize information operations generated by AI and that gaps in ICRC interpretive guidance for information with no evident attribution source and standards for verifying content are the most pressing gaps this sample presents to IHL

Interpersona An International Journal on Personal RelationshipsVol. 20(4)
International Law and Human Rights
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon