While terrorist hostage-taking is uncommon, ideologically motivated hostage-takers often provide a video demonstrating proof of control and proof of life (POL). These videos establish that a hostage is alive at the time the video was made, and hence provide valuable information for families and governments charged with their safe recovery.

What they don’t reveal, however, is the likely fate of hostages.

Theories have been suggested that hostage-taking is a form of organized crime where economic concessions are demanded from rational actors requiring a benefit. While this may apply in many cases, this article argues that each case and hostage-taking group differs, and variables apply dependent on the political situation, the group dynamic and evolving methodology, and the political ‘win’ current to the prevailing conditions (e.g. civil war breakout).

There is limited research on POL videos and narrative exploration, which can inform thinking as to whether hostage-takers are more inclined – through their behaviors and narrative – to execute hostages, or whether they are holding hostages to obtain political concessions to pursue, fund, and finance their strategic attacks and initiatives.

Addressing limited research in this area can help support our understanding of the terrorist group dynamic, its strategic thinking, and the business model it is employing.

Furthermore, any hostage rescue operation can significantly raise the risk to life of a kidnapped hostage held by a terrorist group, so insights gained from analyzing POL videos have the potential to improve operational planning and save lives. In essence, these insights can be invaluable for security services who analyze this footage and try to understand what it means and predict outcomes, negotiators (both government and private security firms) and for security professionals who are considering deployment of staff to high risk, low infrastructure regions.


Three types of hostage-taking

Hijacking – Hostages are held under duress but the main goal of the terrorists is something other than killing the hostages, e.g. the Al-Qaeda-inspired 9/11  attacks, where planes were hijacked and flown into iconic buildings.
 Barricade-siege – Terrorists invade a building to kill or hold people hostage. This type of incident inevitably provokes a quicker response from the authorities and results in a stand-off, e.g. the Bataclan theatre siege where Islamic State of Iraq and the Levant (ISIL) inspired terrorists murdered 90 people and injured many more.
 Kidnapping – Different to the first two categories in that it is normally a predetermined activity involving the instigation of demands or concessions by the hostage-takers.



Killed or released?

To determine what insights POL videos might hold about the fate of hostages, I carried out a coding and content analysis of 51 POL videos involving 62 hostages (predominantly males aged from 23 to 70 years) kidnapped by international terrorist organizations. The cases were recorded from 2003 to 2017.

Table 1 (below) provides a summary of the initial capture. Table 2 (below) summarizes the narrative analysis of the POL videos.


Whom demands are directed at

Who speaks in the video

Hostage-takers shown

Male – 87%

Government – 92% 

Hostage – 67%


Female – 13% 

Family – 6%

Hostage-taker – 6%

(74% died)


Unknown – 2%

Both – 27%


Ransom demanded

Threats made

Deadline given

Weapon pointed







(100% died)


 Table 1. Descriptive statistics derived from the analysis of POL videos

Religion mentioned

Religious symbols

Qu’ran cited 





On a bed – 2%




Standing – 14%




Sitting – 70%




Kneeling – 14%

Other content

Armed or masked captor

Criticism of government

Directed narrative






(75% killed)



 Table 2. Video Content

From the analysis of the sample, hostages taken in Iraq and Syria had the highest death rate in terms of this sample – 78% and 70% respectively. Hostages kidnapped in Afghanistan and Mali had no fatalities (however, it is noted that hostages are still being held at this time in Mali and, therefore, their fate is unknown).
Furthermore, regarding Afghanistan and Mali, these cases accounted for the longest periods held in captivity (five years and five-and-a-half years respectively). This suggests that the longer a hostage is held, the more likely they are of being released.

The POL videos varied in length from 12 seconds to almost 19 minutes. The longer videos tended to involve a much more advanced production, incorporating propaganda and news footage from a political dimension.

In terms of outcome, there was no difference whether the hostages were shown alone or together (in cases where multiple hostages were taken). Fourteen percent of the sample in the POL videos were shown as on their knees in a pleading state. Of note, all hostages in these videos were killed.

JNIM in Mali were noted to have significantly improved their POL productions in the last six years to a very professional output. It is unknown whether this is a strategic imperative or could be answered by access to better equipment and applications that are now available.
It was rare to see hostage-takers delivering a narrative alone and without the hostage in view. Only two cases from the sample fitted this category. They were both ISIS videos from Syria cases and depicted the same hostage-taker. Both cases ended in death39% of the sample analyzed involved armed and masked hostage-takers being depicted in the videos. Of those, 75% ended in the death of the hostages.

Half of the videos involved at least one hostage being killed.



The results identify a causal relationship albeit a significant positive relationship. It should be apparent to the reader that this will be taken into context within the whole context of the investigation and circumstances at that particular time. Political relationships and snapshots in time play an important part of any response mechanisms as well.

The content analysis characteristics add to the understanding of investigators. Demands were predominantly directed at governments (92%) and were political in nature. Significant amounts of demands were for the swap of prisoners. Therefore, there is a high likelihood that a terrorist organization will make demands towards governments. A ¼ of the demands also incorporated a ransom of a significant sum (i.e. millions). Two points are worthy of note regarding the payment of ransoms. Firstly, they are generally never paid at the opening offer and will be reduced by negotiation. Secondly, all of these demands are in the region of what is known in the trade as an ‘unrealistic demand’; that is one where only a government would be able to pay due to the size of the demand. Empirical evidence suggests that both sides know it is illegal to pay ransoms to terrorist groups in any event, so it is possible that the hostage takers’ intention is to demand an unrealistic amount and then use this as justification to kill. This tends to suggest that a realistic demand (i.e. money) would probably signal an intent to release the hostage whereas something totally unrealistic (i.e. US out of Iraq) would lead to death.

Positioning was another factor of note. All POL videos examined where the hostage was filmed kneeling were killed. This would tend to suggest that where a POL video is shown and the hostage is filmed on their knees (and in a pleading, submissive manner), that the outcome may be more likely to be death. 

The majority of videos showed the hostages alone. Where the hostage-takers were seen in the videos then there was found to be a higher likelihood of death as the outcome. 

The study did find a relationship between death as an outcome with the hostage taker(s) shown in the video, weapons pointed at the hostages, deadlines being given (for the demands), threats made and the wearing of orange jump suits. Again, a significant proportion of the video containing hostages wearing orange jumpsuits were from the ISIS videos, with one further in an Iraqi POL video.