MGIMO, deepfakes and military coups

Four years after Lieutenant Obiang's attempted coup in Libreville, the Moscow State Institute of International Relations (MGIMO) has published an article by Roman Kazorin and Andrey Navolokin dedicated to the fight against disinformation and deepfakes. To illustrate the impact that deepfakes can have, they quote one of the articles I published in early 2021 on Lieutenant Obiang's attempted coup in Libreville, considering that according to my article, the video of Ali Bongo's speech just before the coup was a deepfake, and that this had contributed to sparking the coup. But, contrary to what Kazorin and Navolokin claim, the fact is that my article describes exactly the opposite.

In summary, after a long media absence following a stroke, Ali Bongo appeared on Gabonese television on December 31, 2018 to present his New Year's greetings. On January 3, 2019, the GabonReview website announced that members of the Gabonese opposition considered the video of Ali Bongo's speech to be a deepfake and that the Gabonese president had died. On January 7, Lieutenant Obiang attempted a coup. Subsequently, several articles analyzed the link between this video, sometimes presented as a deepfake, and the coup attempt, including an article published on Jeune Afrique, stating: “To justify their act, the putschists in particular affirmed not to have been convinced by the new year presidential video”.

My article therefore stated that, first, the video was not a deepfake1, and, second, that Lieutenant Obiang had not attempted a coup because the video was a deepfake. The video of Lieutenant Obiang's speech, which Kazorin and Navolokin may not have seen, shows that the puschists did not consider that Ali Bongo had died, or that the video of his greetings was a deepfake, since Obiang claimed that the President was ill or diminished, which is obvious on the video of the greetings, and instrumentalized by actors wanting to monopolize power 2.

Furthermore, Obiang's intervention helps us to understand why the coup failed: based on the work of Naunihal Singh, third-order cindynic models have made it possible to describe the role of power perception dynamics in coup attempts. Obiang called on the civilian population to support him, showing Gabonese officers his weakness: this meant that if officers initially hesitated, they subsequently had no choice but to choose the loyalist camp

This case illustrated the value of Relativized Cindynics 3, whereas macro-quantitative approaches to coup prediction had failed to predict Lieutenant Obiang's coup. That said, the need to supplement these unreliable approaches with ad hoc analyses didn't seem to interest the academic or institutional spheres. Not long afterwards, a series of coups that nobody seems to have seen coming overturned the Sahel: in Mali in August 2020 and May 2021, in Burkina-Faso in January and September 2022, in Niger in July 2023. In France, it is claimed that the DGSE's inability to foresee the coup in Niger (or even those in Mali) led the Elysée to replace its Director, Bernard Emié, with that of the DGSI, Nicolas Lerner. Concerning Relativized Cindynics, it may be of interest to note that the development of third-order modelling makes it possible to describe a coup-prevention mechanism within ECOWAS.