Sudan updates from AI Botmaster Jeremy Lichtman of April 25, 2025

How many conflict deaths will there be in Sudan in 2025?

Jeremy’s most recent version of his Multi-AI Oracle is a side competitor with the machine forecasting bots of the VIEWS competition.

Here’s what Jeremy‘s latest version of his Multi-AI Oracle predicts: (Carolyn’s note: Both Jeremy and I agree, these probabilities are, so far, just an experiment. It is too early to know how good these will turn out to be)

Model value:
* Less than 1000: 1%
* Between 1000 and 3000: 2%
* Between 3000 and 5000: 3%
* Between 5000 and 8000: 5%
* Between 8000 and 12000: 20%
* More than 12000: 69%
Here’s how Jeremy’s Multi-AI Oracle says it arrived at these probabilities:

Obtaining outside data.
Checking internet sources, using keywords: Sudan conflict deaths 2025
Results from around the web: {think}

Okay, let’s tackle this query step by step. The user wants current web results about Sudan conflict deaths in 2025, with a summary, tables of relevant data, and a time series if possible. They also want URLs listed at the end.

First, I need to check the search results provided. Let’s go through each one:
1. **ACLED (Search result 1 and 5):** They report over 28,700 fatalities by November 2024, including 7,500 civilians. They mention that this is likely an underestimate, with some sources suggesting up to 150,000 deaths. Also, 30% of the population displaced.
2. **Science.org (Result 2):** Similar numbers to ACLED—28,700 deaths, 7,500 from intentional injuries. Highlights that starvation and disease might be bigger killers than direct violence. Mentions both SAF and RSF using starvation as a weapon.
3. **CFR (Result 3):** Cites the former U.S. envoy’s estimate of up to 150,000 killed since the start, aligning with the higher end from other sources.
4. **Genocide Watch (Result 4):** Claims 61,000 killed in Khartoum alone, with 26,000 direct war deaths. Notes that 90% of deaths in Khartoum are unrecorded, leading to severe undercounting. LeMonde’s estimate of 150,000 civilian deaths is mentioned here too.

Now, the user wants tables. Let’s see what data points we have. The main numbers are the reported fatalities from ACLED (28,700) and the higher estimates (150,000). There’s also data on displacement (over 14 million, 30% of the population). However, there’s no time-series data provided in the search results. The latest dates mentioned are up to December 2024 and January 2025, but no monthly breakdowns. So a time series table might not be feasible. Instead, a comparison table of different estimates would make sense.

The summary should highlight the range of estimates, sources of undercounting, and contributing factors like blockade of aid. Also, note the geographic spread (Khartoum, Darfur) and the actors involved (SAF, RSF).
For URLs, the sources used are the ones provided: acleddata.com, science.org, cfr.org, genocidewatch.com. Need to list all these at the end, excluding any duplicates.

Make sure to cite each source correctly in the summary. For example, when mentioning ACLED’s numbers, cite [1][5]. When talking about higher estimates, cite [3][4]. The starvation tactics would be from [2][4].

Check if any tables can be created. The main data points are:
– ACLED: 28,700 reported fatalities (direct), 7,500 civilians.
– CFR/Genocide Watch: Up to 150,000 total deaths.
– Khartoum deaths: 61,000 total, 26,000 direct.

So a table comparing these estimates would be useful. Also, a table on displacement and humanitarian impact.
Need to structure the answer with headers but start with a summary. Follow the formatting rules: no headers at the start, use ## for sections, bold subsections if needed. Use tables where comparisons are made. Ensure citations are in brackets after each relevant sentence.

Finally, conclude with the list of URLs. Make sure they’re correctly formatted and only include the ones actually cited. The ongoing conflict in Sudan has resulted in catastrophic human losses, with widely varying estimates of fatalities due to challenges in data collection and underreporting. Key sources provide the following insights:
—
## Reported Fatalities and Estimates
– **ACLED data** (through November 2024):
– **28,700+ reported fatalities**, including **7,500 civilians killed in direct attacks** [1][5].
– These figures are considered a significant undercount, with displacement affecting **over 30% of Sudan’s population** (14 million people) [1][2].
– **Higher estimates**:
– The former U.S. envoy for Sudan and *LeMonde* suggest **up to 150,000 total deaths** due to combined effects of violence, starvation, and disease [3][4].
– In Khartoum alone, **61,000 deaths** have been reported, with **26,000 directly tied to combat** [4].
—
## Humanitarian Impact
**Key factors exacerbating casualties**:
– **Starvation as a weapon**: Both the Sudanese Armed Forces (SAF) and Rapid Support Forces (RSF) have blockaded food and medical aid, affecting **30 million people** in need [2][4].
– **Infrastructure destruction**: Hospitals, schools, and utilities have been targeted, compounding deaths from preventable causes [2][4].
– **Ethnic violence**: The RSF has been accused of genocide against non-Arab groups in Darfur, echoing earlier atrocities [2][4].
—
## Data Comparison Table

## Challenges in Data Accuracy
– **Underreporting**: 90% of deaths in Khartoum go unrecorded due to collapsed infrastructure and targeting of aid workers [4].
– **Methodological limits**: Satellite imagery and modeling are used to estimate casualties, but gaps persist [2].
—
URLs of sources used:
1. https://acleddata.com/conflict-watchlist-2025/sudan/
2. https://www.science.org/content/article/how-many-have-died-sudan-s-civil-war-satellite-images-and-models-offer-clues
3. https://www.cfr.org/global-conflict-tracker/conflict/power-struggle-sudan
4. https://www.genocidewatch.com/single-post/sudan-genocide-emergency-january-2025
5. https://acleddata.com/africa/horn-of-africa/sudan/

Couldn’t get base forecast

Querying Claude (AI predicts: [“Less than 1000”: 0.01, “Between 1000 and 3000”: 0.02, “Between 3000 and 5000”: 0.05, “Between 5000 and 8000”: 0.12, “Between 8000 and 12000”: 0.25, “More than 12000”: 0.55] – confidence: 6)

Querying Mistral (AI predicts: [“Less than 1000”: 0.01, “Between 1000 and 3000”: 0.02, “Between 3000 and 5000”: 0.03, “Between 5000 and 8000”: 0.04, “Between 8000 and 12000”: 0.2, “More than 12000”: 0.7] – confidence: 7)

Querying OpenAI (AI predicts: [“Less than 1000”: 0.01, “Between 1000 and 3000”: 0.01, “Between 3000 and 5000”: 0.03, “Between 5000 and 8000”: 0.05, “Between 8000 and 12000”: 0.15, “More than 12000”: 0.75] – confidence: 6)

Question Type: Multiple Choice
Confidence: 6
MAPD: Avg: 0.045555555555556, Max: 0.13333333333333
# LLM responses: 3

Model value:
* Less than 1000: 1%
* Between 1000 and 3000: 2%
* Between 3000 and 5000: 3%
* Between 5000 and 8000: 5%
* Between 8000 and 12000: 20%
* More than 12000: 69%

The ongoing conflict in Sudan has led to significant humanitarian and fatal consequences, with estimates of fatalities ranging from over 28,700 to as high as 150,000 by late 2024. The conflict involves the Sudanese Armed Forces (SAF) and the Rapid Support Forces (RSF), with reports of both sides employing tactics such as starvation and targeting civilians, further exacerbating the crisis. The humanitarian situation remains dire, with 30 million individuals in need of aid and 14 million displaced. Critical infrastructure destruction and aid blockades suggest mortality rates will likely continue to be high through 2025. However, the actual death toll is uncertain due to underreporting and challenges in data collection, particularly in areas like Khartoum. Despite this grim outlook, factors such as international pressure, potential peace agreements, or improved humanitarian access could mitigate future fatalities. Additionally, the unpredictability of conflict scenarios and dynamic socio-political developments could significantly impact these projections.

Past forecasts on Sudan’s conflict fatalities —>

Guide to Multi-AI Oracle’s Reports (some of these only apply to older versions)
Bayesian calculation: a mathematical technique that helps determine the conditional probability of an event based on prior knowledge and new evidence.
MAPD: Mean Absolute Percentage Deviation calculates the average percentage difference between actual values and predicted values, providing a relative measure of error.
Median: We collect the output values from the LLMs in an array. We then take the median value. This is between 0 and 1 (i.e. 0.15 = 15%).
Base rate: Currently we take OpenAI’s value as the base rate. We are working on a better value using a different prompt that takes into account process. The base rate helps to determine if the median makes sense.
SD: The standard deviation between the median and the base rate.
Confidence: We query each of the LLMs on how confident they are of their predictions (between 0 and 10) and take a median. Because the LLMs tend to be overconfident, we take anything lower than 6 as being low confidence. This factors into the overall model.
Conf Mode: Based on the confidence value. >=9 is high confidence. Below 6 is low confidence (this is also triggered by an exceptionally high SD).
Mellers: This refers to Barbara Mellers, specifically a paper she wrote that includes a formula for moving values towards an extreme (i.e. 0 or 1).
Reverse Mellers: This uses the formula from above, but with a sub-1 coefficient to move the values closer to 50%.
Theory of Mind: We ask the LLMs what they think other LLMs would predict. We hope that this makes them consider the questions more deeply.
Beta Distribution: Currently unused, but possibly of interest. This is based on the median, the base rate and the SD.
Close Type: We noticed that the appropriate base case for some questions is closer to the extremes, while others are closer to 50%. When we have a low confidence value, this helps us to determine whether to extremize or de-extremize the value. ‘A’ implies closer to zero. ‘B’ implies closer to 50%. ‘C’ implies closer to 100%.
# LLM Responses: The count of LLMs that responded. We query 5 currently, but they have a relatively high API failure rate.
Model value: For normal confidence mode, this is the median. For high confidence, this is the Mellers value. For low confidence, this is either the Mellers or Revers

Background on VIEWS: Near the end of 2024, the latest forecasting competition of the Violence and Impacts Early-Warning System (VIEWS) was launched. On Dec. 4, 2024, during a telecon, a group of machine forecasting experts involved in VIEWS observed Jeremy’s demonstration of his Multi-AI Oracle. Although it was too late for Jeremy to sign up as a VIEWS competitor, they urged him to run a side competition for everyone’s benefit. Jeremy chose their Sudan conflict deaths question.

Background on the Sudan conflict, as evaluated by the team of Jeremy Lichtman and Carolyn Meinel:
(1) The importance of the Sahel, a region in Africa that includes Sudan:
Why the Sahel Actually Matters (Possibly To You, Too)
(2) April 29, 2024: Today, a Russian diplomat is negotiating gold mining with the Sudanese Minister of Minerals.
(3) June 3, 2024: Today the African Union Announced Plans to Mediate in Sudan’s Civil War
(4) Sept. 11, 2024 Over 100K starvation deaths 96% likely in Sudan by 1 Jan 2025 Below, map of control in Sudan’s civil war as of April 15, 2025.

Below, before and after pictures of the devastation of the civil war in Sudan.

https://www.alnilin.com/
https://www.alrakoba.net/
https://www.altaghyeer.info/ar/
https://www.altaghyeer.info/en/
https://www.alttahrer.com/
https://www.alzaeemsd.com/
https://www.assayha.net
https://www.cover-sd.com/
https://kushnews.net/
https://rakobanews.com
https://www.sudanakhbar.com/
https://sudanjem.com/
https://suna-sd.net/en
https://sudantimes.net/
https://www.sudantribune.net/
https://www.sudantribune.com/
https://sudaneseonline.com/
https://www.sudanile.com/
https://sudanow-magazine.net/
https://www.sudaress.com/
https://www.tagpress.net/