Experimental Red Teaming to Support Integration of Information in Joint Operations

Authors: Gary Ackerman, Douglas Clifford, Anna Wetzel, Jenna LaTourette and Hayley Peterson (CART)

Publication Preview

This report supported SMA’s Integrating Information in Joint Operations (IIJO) project. For additional speaker sessions and project publications, please visit the IIJO project page.

As part of the modeling and simulation phase of the IIJO effort, the Center for Advanced Red
Teaming (CART) worked closely with the ICONS Project at the University of Maryland to
employ two separate yet integrated human simulation approaches to test and build on the
findings of earlier components of the IIJO effort. The CART portion of the simulation
involved:

1. Distilling 46 propositions from the Net Assessment and TTXs into 12 explorable insights (EIs) regarding the competitive information environment.
2. Testing these EIs in six scenario-based Red Team experiments using 223 U.S.-based proxy participants from similar cultural backgrounds to actual adversary target populations (Taiwan for the Asian context and several Southeast European countries for the European region).
3. Collecting data on several measures of messaging effectiveness and analyzing this data to validate or shed new light on the EIs. The experiments yielded a number of takeaways relevant to the IIJO project

The experiments yielded a number of takeaways relevant to the IIJO project.

1. The United States begins with a reputational / perception advantage over its GPC competitors, but the gap is fairly narrow between the U.S. and China in the European context.
2. Wherever it was tested (Scenarios 2, 4, and 5), there was robust evidence that GPC adversary propaganda that seeks to cast the United States in a negative light is effective in lowering attitudes towards the U.S., trust in the U.S. and U.S. influence among targeted audiences in non-GPC states (in our experiments, Taiwan and the states of southeastern Europe). Given that the perceptions of these audiences can impact their own countries’ and others’ military and political support for the United
   States, these findings confirm that IIJO is crucial and that the United States military needs to place significant emphasis on OIE moving forward. Furthermore, messaging to foreign populations (whether through traditional or new channels such as social media) cannot be left out of operations.
3. What countries do (as opposed to only what they say) matters. Hypocrisy by any GPC state leads to negative perceptions among target audiences, but there is no evidence that this harms the U.S. more than its GPC adversaries.
4. There is some, although not robust, evidence to indicate that messaging regarding U.S. economic success may not go as planned and could actually hurt foreign perceptions of the United States, whereas the jury is still out on the effects of similar messaging for other GPC states.
5. The effectiveness of messaging regarding the economic shortcomings of GPC adversaries like the PRC remains unclear.
6. There is no experimental evidence to suggest that emphasizing U.S. values is advantageous in its messaging and at least a possibility that doing so might negatively affect perceptions about the U.S.
7. Adopting a victimization narrative does not appear to be an effective messaging strategy in OIE, and may in fact backfire, lowering the believability of the message
   and perceptions of the country utilizing this approach.
8. There is insufficient evidence to suggest that a non-U.S. government messenger is preferable to a U.S. government messenger, but more research is required on this point.
9. There was partial support (but only in one AOR) for the proposition that uncrafted and untargeted messages are more effective in influencing perceptions about the U.S., but further research is required to determined when this finding is applicable and when it is not.
10. With the possible exception of how much the message is believed and shared, there was no experimental support for the notion that positive, proactive messages
    are more effective than negative, reactive messages.
11. There is some limited, provisional support for the proposition that adversary messages that attack common values between the U.S. and the target population will have a more powerful (negative) effect.
12. The proposition that messaging that resonates with current beliefs and perceptions of the target audience will have greater believability received no experimental
    support. There were inconclusive findings as to the effects of more resonant messaging on other measures of effectiveness, with contradictory findings in the Asian and European samples.
13. There is evidence that, in a crisis, it is better to send no message than to urge allies to refrain from escalation.
14. There can sometimes be unforeseen effects to OIE. For example, in some experiments, messaging focused on one country actually affected perceptions of other
    GPC states (including the state doing the messaging).

In addition to the substantive findings, at the programmatic level, the experiments demonstrated how human simulation can be used to test emerging phenomena or novel ideas that arise from the insights of experts and various other knowledge artifacts developed during the course of a typical SMA study. By exposing these insights to realistic simulations involving disinterested participants at scale, the use of an integrated human simulation approach (experiments plus table-top exercises) can both validate previous findings and reveal new dynamics in complex systems like the OIE.

‍