Did the M23 commit war crimes as claimed by Amnesty International, and does it target Hutus as claimed by Human Rights Watch ?

by 27 August 2025Fact-checking

In two reports published last Wednesday, Amnesty International and Human Rights Watch both accused the M23 rebels of committing war crimes. An emergency meeting of the UN Security Council was held on Friday. African Facts revisits these accusations and their reliability.

Fighs are currently taking place in the Virunga Mountains in eastern Democratic Republic of Congo opposing the Congolese March 23 Movement (M23), which is supported by neighbouring Uganda and Rwanda, to the Democratic Forces for the Liberation of Rwanda (FDLR), an armed group formed by perpetrators of the 1994 Rwandan genocide against the Tutsi; and the Nyatura of the Coalition of Movements for Change (CMC), a Congolese Hutu militia affiliated with the FDLR. Since the beginning of the year, the clashes have been concentrated in the north of the Rutshuru territory, North Kivu province. Over the past few months, civilians have died in the fighting, particularly in Binza, as our correspondent on the ground has previously reported.

Against this backdrop, since late July 2025, activists close to the Nyatura-CMC have been disseminating allegations that the M23 is responsible for massacres and summary executions of Hutu residents of Binza. Last Wednesday, Human Rights Watch and Amnesty International published two reports accusing the M23 of war crimes. On Friday, the United Nations Security Council held an emergency meeting regarding the situation in eastern DRC.

However, a thorough analysis of the Human Rights Watch and Amnesty International reports reveals significant methodological issues and substantial gaps in demonstrating the alleged facts that form the basis of the serious war crime accusations made by the two NGOs. HRW partly adopts the rhetoric of one belligerent and emphasises the ethnic affiliation of the alleged victims. Furthermore, it appears that only 8.5% of the specific facts or particular events mentioned by Amnesty International are documented or corroborated by at least three sources. Neither organisation conducted fieldwork.

African Fact had reported on the initial accusations against the M23 at the beginning of this media and diplomatic sequence. On 31 July, Reuters published a story claiming that the M23 had massacred 169 “Hutu farmers” over the previous month. Reuters obtained this information from the United Nations Joint Human Rights Office in the DRC, who themselves based it on information provided by a “Collective of Victims of Rwandan Aggression” and a “Rutshuru Territorial Youth Council”. These two organisations are intricately linked and comprise Nyatura-CMC militiamen. On 6 August, the United Nations High Commissioner for Human Rights claimed that 319 civilians had been “killed by the M23”. In total, five different accounts of these alleged crimes have emerged in the space of two weeks, to which Human Rights Watch has added a sixth today.

This graph shows the number of alleged victims in reports, articles, and statements between July 25 and August 22, 2025.

Congolese government authorities are involved in spreading these accusations. On 12 August, the FARDC spokesperson released a filmed statement in which he claimed that “targeted massacres of defenceless Congolese civilian populations, mostly Hutu, [are taking place] in North Kivu”. At the emergency meeting of the Security Council on 22 August, the Kinshasa representative seized the opportunity to double down. “These crimes have a manifest ethnic dimension. Civilians, particularly Hutu people, have been targeted solely because of their ethnic affiliation. It is a deliberate attempt to reshape the Congolese social fabric through violence”, said the Congolese ambassador, who repeatedly spoke of “transplanting of unidentified populations” in areas under M23 control.

For several years, the Kinshasa authorities and their allies have openly relied on racist hatred in their public statements in an attempt to unite and mobilise the population against an alleged “internal enemy” who is portrayed as both powerful and elusive: Congolese Tutsis, who are supposedly “infiltrators” conspiring to bring about the defeat and division of the country. This xenophobic rhetoric notably includes the fantasy of a “replacement” project of groups considered “indigenous” as opposed to the Tutsi, who are designated as allogenic.

By suggesting the possibility of an “ethnic cleansing” of the Hutu in its latest report, Human Rights Watch runs the risk of aligning itself with this narrative and lending it significant credibility.

An “ethnic cleansing” of the Hutu in Rutshuru ?

Human Rights Watch claims that the M23 “summarily executed over 140 civilians” and says it interviewed 25 witnesses in support of these accusations. However, in its publication, the NGO only cites three anonymous testimonies from individuals who claim to have directly witnessed 56 murders committed by the M23. One witness is a woman who claims to have been present during the execution of 47 people, while the other two are men who claim to have seen members of their families being killed. Three other murders mentioned by HRW were reportedly recounted to one of its sources by another witness, constituting hearsay. Another death was reported to HRW by an unspecified number of anonymous sources who found a body; this gives no indication as to who committed the crime. Therefore, there remain 81 deaths for which the organisation provides no explanation.

The NGO claims to have “analyzed relevant videos and photographs” corroborating the accusations it makes. However, the organisation later acknowledges in its report that it “was unable to confirm the location, time, or dates the images were taken”. Without context or cross-referencing regarding the circumstances of these deaths, these elements cannot therefore be considered relevant. Did HRW nevertheless include the corpses visible in these images in its macabre count ? The authors of the report did not wish to answer this question.

This graph illustrates the breakdown of the 141 alleged deaths cited in the Human Rights Watch report, categorised by the number and type of sources: hearsay only; a single anonymous source; an indefinite number of witnesses; or no source cited.
Répartition des morts allégués par type et nombre de sources (HRW)
Type / robustness of sourceAlleged number of deaths
Hearsay only3
1 andonly anonymous source56
Indefinite number of witnesses1
No source cited81
Total141

Human Rights Watch swiftly stated that “most victims were ethnic Hutu” and then specified that there were “targeting of Hutu civilians”. It is unclear how the NGO determines the ethnic affiliation it assign to the victims of the crimes it denounce, or what evidence they rely on to conclude that such targeting is taking place. HRW also emphasises the alleged use of machetes by the M23 to carry out what its Nairobi-based researcher describes as “dozens of summary executions of primarily Hutu civilians”.

These accusations echo those in a Reuters article from 31 July, which claimed that “Hutu farmers” had been specifically “targeted” by the M23, whose fighters had committed their crimes “using machetes”. But Human Rights Watch goes further, stating that this “raises grave concerns of ethnic cleansing”.

The Human Rights Watch report is thus part of a media and diplomatic sequence that began a month earlier with the Reuters article and was followed by statements from the United Nations High Commissioner for Human Rights. It is worth noting that the dates, locations and circumstances of the events mentioned by HRW in its report correspond to allegations made by organisations and individuals associated with the Nyatura and FDLR. Similar accusations were made by United Nations officials two weeks earlier, to whom these same actors had provided the supposed sources to corroborate them. However, HRW does not provide any additional evidence to support these accusations.

It is difficult not to suspect the influence of the same activists close to the FDLR and the Nyaturas in the accusations of ethnic targeting and the use of machetes made by Reuters and then Human Rights Watch against the M23. In the background, we can detect the ingredients of the negationist thesis of the ‘double genocide’ imagined by the Rwandan genocidaires and their supporters immediately after the 1994 genocide, and repeated tirelessly ever since. According to this thesis, the genocide of the Tutsis in Rwanda was allegedly followed by a genocide of the Hutus in the Congo. This is an accusation in a mirror, turning victims into executioners and, through a fallacious dialectical reversal, relativising the facts and re-establishing a biased – and racist – representation of reality among foreign observers : in the depths of Africa, two ethnic groups, driven by an atavistic rivalry, would be killing each other with machetes.

In an attempt to better understand the situation, African Facts contacted the researcher who wrote the Human Rights Watch report. She directed us to the Africa director of the NGO, who refused to discuss it over the phone and demanded written questions. We sent these to him. He did not respond. (NB: Lewis Mudge sent us his answers in writing two hours after the article was published. You can find these at the bottom of the page.)

Other war crimes accusations

African Facts has also carefully read the Amnesty International report, which was published on the same day as the Human Rights Watch report. The NGO does not repeat the allegations concerning supposed massacres that allegedly took place in the Binza area. On the other hand, it does accuse the M23 of rape, arbitrary executions, physical violence, kidnapping, attacking hospitals and human rights defenders.

We counted 47 specific facts or particular events mentioned in the body of the document that correspond to alleged M23 crimes. We paid particular attention to the sources on which the organisation bases its information, which it honourably mentions systematically.

In 31 cases Amnesty International only provides a single anonymous testimony in support of each particular event or specific fact mentioned. In order to present factual truths to the public, rigour normally requires cross-referencing a plurality of independent sources for each event or situation described. This is clearly not the case here.

Furthermore, some of the testimonies presented by the NGO were reported to the organisation, but were not collected from the individuals who allegedly provided them. Others are presented as “interviews with a person familiar with the incident” or even a “confidential source”. In each of these cases, it is therefore hearsay. Such statements are by nature subject to caution and cannot stand on their own. And, no less than six facts presented by the NGO are based solely on this type of source.

This chart breaks down the 47 alleged facts in the Amnesty International report according to the number and type of sources on which they rely. These sources may be based on hearsay only, or on one, two or three anonymous sources. Alternatively, they may be based on material evidence.
Number of facts by type of sources
Type of sourceNumber of facts
Hearsay only6
1 and only anonymous source31
2 anonymous sources6
3 anonymous sources1
Material evidence3

Of the facts alleged by Amnesty International, only one is based on three sources, even though these remain anonymous and one of them is indirect. Two cases of violence and one murder are documented by videos. Therefore, beyond the cumulative effect produced by the NGO’s argumentation and the emotion it evokes, only 4 out of 47 facts are based on a minimally acceptable amount of evidence. This represents 8.5% of the facts mentioned in the report.

This graph illustrates the robustness of the sources that corroborate each fact within each major category of accusation, according to whether the source is hearsay, one source, two sources, three sources or material evidence.
Robustness of facts based on the type of source
Type of facts Hearsay only 1 source 2 sources 3 sources Material evidence
Attacks on HRD219000
Rapes16200
Attacks on hospitals33210
Summary executions and murder12101
Physical violences00002
Kidnappings of young men01100

Upon reading the report, one cannot help but question Amnesty International’s use of verbs in the present or past indicative, given that the conditional should be used wherever there is a lack of material evidence or corroborating sources to establish precise, detailed and contextualised facts. In terms of form, this is far from the cautious formulations usually seen in this type of document.

African Facts contacted Amnesty International to try to understand this. “I don’t think this framing is correct”, responded Christian Rumu, who participated in the NGO’s advocacy campaign alongside the release of the document. “We didn’t discuss every event in depth, and I don’t want to provide more detail for fear of exposing the people who talk to us. I believe our methodology is quite clear in the report. Your desire to understand how we arrived at a conclusion for each incident does not correspond to our way of doing things. As for how we reach conclusions or fact-check these incidents, I would really refrain from venturing into that”, he evaded.

In our opinion, all the NGO’s allegations must consequently be considered with extreme caution.

The reliability of the sources and researchers into question

Neither Human Rights Watch nor Amnesty International went to the field, nor did they meet any of the cited witnesses in person. This is another major weakness that both investigations have in common. Clearly, these investigations were conducted thousands of kilometres away from the locations where the alleged events took place.

Furthermore, all of the testimonies produced by the two organisations are anonymous. The anonymisation of witnesses is justified in order to protect sources. However, basing such serious charges as war crimes or allegations such as ‘ethnic cleansing’ solely on anonymous sources, with no material evidence or identified sources to corroborate them, seems particularly unwise.

This diagram assesses the reliability of the actors who produced the data, based on seven criteria. Hover over an actor in the legend to highlight them.

Direct access to the field and witnesses, cross-checking of facts, absence of visible ideological bias, non-belligerence, methodological transparency and verifiability – as much as possible – of claims are all objective criteria that should underpin the work of human rights organisations and journalists in all circumstances. This is all the more important when armed conflicts are combined with an information war.

But in these mountains at the heart of Africa, which some people seem intent on portraying as distant and shrouded in darkness, these basic principles do not seem to apply. Narratives are shaped, and good and bad roles are assigned, from the outside with no regard for the facts on the ground. During the period covered by the Amnesty International report, at least three civilians were lynched and eaten by pro-government militia in the concerned area. The NGO doesn’t say a word about it.

27/08/2025 | Fact-checking

Methodology

The nebulous use of sources by the two NGOs has complicated the analysis of the two reports.

We chose only to count people as witnesses if they were explicitly presented as such, except in cases where their accounts suggested that they had not directly witnessed the crimes. We classified as hearsay the statements of individuals who reported to the NGOs the testimonies of other witnesses, as well as individuals cited by Amnesty International in its footnotes as “familiar with the incident” or “confidential sources”, whereas all others were cited as ‘witnesses’ (in the french version).

Each criminal act and each particular event with a unity of time and place was counted as a precise fact.

Assigning a ‘score’ in the reliability assessment aims to make the various points addressed in our two fact-checking articles more tangible and visualisable in a diagram, and to put them into perspective.

We did not award the United Nations Human Rights Office in the DRC all the points in the field presence, direct access to witnesses and non-belligerence criteria. Although the United Nations Organisation Mission in the Congo is present in North Kivu, it is not deployed in the Binza area and therefore does not have direct access to witnesses. Furthermore, despite being a peacekeeping mission, it is an armed force in a war zone. This is what we wanted to illustrate.

Similarly, we did not award all the points to HRW in the absence of ideological bias criterion for the reasons set out in the article.

We established three sub-criteria regarding methodological transparency:
systematic and clear mention of the elements that establish the facts; rigour in demonstrating, taking all the belligerents into account.

 

Addendum :

Lewis Mudge explained that HRW’s methodology consisted of citing only “some testimonies to illustrate certain points”. According to him: “the testimonies quoted in this report only represent a fraction of the cases we documented, we have all of them on file from this case”. HRW would only use single-source testimony on “very rare occasions”. Regarding the photos and videos mentioned in the report, Lewis Mudge responded: “we are being honest that we are unable to confirm certain aspects of the videos and photos even though we lend credence to them based on their sourcing”. He says he determined the ethnicity of the alleged victims after he “asked family members and people close to the victims”. When asked about the evidence on which the NGO bases its conclusion of ethnic targeting, he states that “the scale and geographical scope of the killings, alongside the fact that most of the victims were Hutu, are some of the factors considered when raising these concerns”.

When asked about the risk of his report being exploited by xenophobic or conspirationnist activists, Lewis Mudge responds by stating that he reports “on the facts around human rights abuse by all actors in the region through an objective lens”. He adds: “I’m sure you can appreciate that we have a responsibility to report verified facts, even if some actors attempt to distort them, because transparency and accuracy are essential to accountability”.

In this case, these actors are actually the source of the information and accusations made by Human Rights Watch. Read our previous fact-checking article.

A new source of reliable information

Many conflicts on the African continent are poorly understood today due to a lack of independent, factual and easily accessible sources of information. In our view, this information gap is a worrying and dangerous vacuum that allows misleading narratives, fanciful figures, fake news and hate speech to flourish. African Facts has therefore made it its mission to fill this which is conducive misinformation.

Stay updated

Receive our upcoming articles in your inbox and stay informed about developments at African Facts.