Sudan updates from AI Botmaster Jeremy Lichtman of May 1, 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:

Model value:
* Less than 1000: 1%
* Between 1000 and 3000: 2%
* Between 3000 and 5000: 7%
* Between 5000 and 8000: 12%
* Between 8000 and 12000: 19%
* More than 12000: 59%

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}

Alright, I need to address the user’s query about the current web results regarding Sudan conflict deaths in 2025. Let me start by reviewing the provided search results to gather the necessary information.

First, looking at the sources:
1. Source [1] from ACLED mentions over 28,700 reported fatalities by November 2024, including 7,500 civilians. They also note that some estimates go up to 150,000 deaths. Displacement is over 30% of the population. This source is from December 2024, so it’s recent but just before 2025.
2. Source [2] from Science.org in February 2025 discusses the challenges in estimating deaths, with figures ranging from 20,000 to 150,000. It highlights that starvation and disease might be causing more deaths than direct violence. This source provides a more scientific perspective and includes data up to early 2025.
3. Source [3] from Genocide Watch in January 2025 cites LeMonde’s estimate of at least 150,000 civilian deaths. It also gives specific numbers for Khartoum and El Fasher, indicating underreporting issues. This is a detailed account focusing on specific regions and atrocities.
4. Source [4] from CFR in March 2025 reports 50 civilians killed in Khartoum recently, showing ongoing violence. This is the most recent data point and indicates that the conflict is still active in 2025.
5. Source [5] from NRC in April 2025 estimates over 150,000 deaths and 14 million displaced. This is the latest figure and aligns with the higher estimates from other sources.

The user wants a summary with tables, including time series data if possible. However, the provided data doesn’t have a detailed time series with periodic values. The closest thing is the ACLED data up to November 2024 and the CFR report from March 2025. The challenge here is that the data spans the end of 2024 and the start of 2025, so I need to clarify the timeline.

I should structure the answer by summarizing the death toll estimates, sources of underreporting, regional breakdowns, and humanitarian impact. Tables can compare the different estimates and possibly the regional data from Khartoum and Darfur.

I need to ensure all sources are cited correctly. For example, the 150,000 figure is mentioned in sources [1], [3], and [5]. The March 2025 incident in Khartoum is from [4]. The breakdown in Khartoum and El Fasher is from [3].
Time series data isn’t explicitly available, but I can mention the progression from late 2024 to early 2025. However, without monthly data, a detailed table might not be feasible. Instead, a table summarizing the different estimates and their sources would be more appropriate.

I should also include the URLs at the end as requested. Make sure each citation is correctly placed after the relevant sentence. Avoid using markdown in the thinking process but in the final answer, structure it with headers and tables as per the format rules.

The Sudan conflict in 2024–2025 remains one of the deadliest ongoing crises globally, with significant discrepancies in reported death tolls and widespread humanitarian suffering. Below is a synthesized summary of available data and key findings from verified sources:
—
### **Reported Death Toll Estimates**

—
### **Regional Breakdown**
**Khartoum**
– Total deaths: ~61,000 (26,000 direct combat deaths) [3].
– 90% of deaths unrecorded due to infrastructure collapse [3].
**Darfur (El Fasher)**
– 782+ civilians killed in indiscriminate shelling (September 2024–January 2025) [3].
– RSF siege tactics exacerbate starvation and displacement [2][3].
—
### **Humanitarian Impact**
– **Displacement**: Over 14 million people displaced (30% of Sudan’s population) [1][5].
– **Starvation as a weapon**: Both SAF and RSF block food/medical aid to civilians [2][3].
– **Infrastructure destroyed**: Hospitals, schools, and utilities targeted [2][4].
—
### **Key Trends and Challenges**
1. **Underreporting**: Limited access for observers and destroyed record-keeping systems inflate uncertainty [1][2][3].
2. **Indirect deaths**: ACLED and Science.org emphasize that starvation, disease, and displacement fatalities likely surpass direct combat deaths [1][2].
3. **Ethnic targeting**: RSF accused of genocide against non-Arab groups in Darfur [2][3].
—
### **Sources**
1. [ACLED Conflict Watchlist 2025: Sudan] (https://acleddata.com/conflict-watchlist-2025/sudan/)
2. [Science.org: How Many Have Died in Sudan’s Civil War?] (https://www.science.org/content/article/how-many-have-died-sudan-s-civil-war-satellite-images-and-models-offer-clues)
3. [Genocide Watch: Sudan Genocide Emergency] (https://www.genocidewatch.com/single-post/sudan-genocide-emergency-january-2025)
4. [CFR Global Conflict Tracker] (https://www.cfr.org/global-conflict-tracker/conflict/power-struggle-sudan)
5. [NRC: Sudan Humanitarian Crisis] (https://www.nrc.no/feature/2025/the-humanitarian-crisis-in-sudan-explained)
Data collection remains fragmented, but these sources collectively underscore a catastrophic humanitarian emergency with escalating violence and systemic underreporting.

OpenAI gives a base rate of probability of death estimates surpassing 12,000 (Carolyn’s note: OpenAI failed to return a value. See below for another way this bot calculates the base rate.)
The question difficulty is rated 7 (0 to 10)
Historical weighted factors include: 0.30.250.20.150.1

A Bayesian calculation could be performed as follows:
Bayesian calculation combines historical factors adjusted by the recent increase in reported deaths. Given the ongoing conflict, the likelihood increases. Calculation: Prior probability (0.8) * Evidence (prolonged conflict and international reports) = Posterior probability of 0.95.
Bayesian base rate: 0.95
Sufficient news to provide a good forecast? 1 (0 or 1)
News is relevant, topical and unbiased? 1 (0 or 1)
Question classification: reference_class
Expected time (days) until question close: 240  (Carolyn’s note: It’s 244 counting today.)

The following were considered in order to produce this base rate:
The base rate is based on historical conflict death estimates in Sudan, with recent reports escalating from 28,700 to estimates as high as 150,000, indicating a significant probability of high death counts.

Ideally, the news feed would contain the following sorts of information for a better forecast:
Additional detailed, on-ground observational reports and validated numbers from neutral parties would aid in accuracy. Satellite data and independent analyses are also critical.

Some potential divergent considerations that might affect the base rate:
Potentially underreported or misclassified deaths could adjust estimates downwards if peace efforts unexpectedly succeed, reducing conflict intensity.

The following chain of events are necessary for the question to resolve positively:
Continued accurate reporting of conflict deaths in Sudan. Moderate – Improvement in humanitarian access and verification on the ground. Low – Stabilization and peace negotiations in Sudan. Low

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”: 0.2, “More than 12000”: 0.6] – confidence: 6)

Querying Mistral (AI predicts: [“Less than 1000”: 0.01, “Between 1000 and 3000”: 0.05, “Between 3000 and 5000”: 0.1, “Between 5000 and 8000”: 0.19, “Between 8000 and 12000”: 0.25, “More than 12000”: 0.4] – confidence: 7)

Querying OpenAI (AI predicts: [“Less than 1000”: 0.01, “Between 1000 and 3000”: 0.02, “Between 3000 and 5000”: 0.07, “Between 5000 and 8000”: 0.1, “Between 8000 and 12000”: 0.15, “More than 12000”: 0.65] – confidence: 7)

Question Type: Multiple Choice
Confidence: 7
MAPD: Avg: 0.057777777777778, Max: 0.16666666666667
# LLM responses: 3

Model value:
* Less than 1000: 1%
* Between 1000 and 3000: 2%
* Between 3000 and 5000: 7%
* Between 5000 and 8000: 12%
* Between 8000 and 12000: 19%
* More than 12000: 59%

The AI analyses focus on the ongoing conflict in Sudan, highlighting a broad range of estimated fatalities from 28,700 to over 150,000 by the end of 2024, with continued violence into 2025. The predictions indicate a high likelihood of exceeding 12,000 deaths in 2025 due to factors like underreporting, indirect deaths from starvation and disease, and systematic infrastructure destruction. Despite this grim outlook, there is also consideration for potential scenarios that could lower the death toll, such as successful peace negotiations, international intervention, and improved humanitarian access. However, the unpredictability of the situation and potential for underreporting or overestimation of deaths remain significant challenges in accurately assessing the situation.

Runtime: 82 seconds.

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/