The Global Risks Perception Survey (GRPS) is the World Economic Forum’s source of original risks data, harnessing the expertise of the Forum’s extensive network of academic, business, government, civil society and thought leaders. Survey responses were collected from 4 September to 9 October 2023 from the World Economic Forum’s multistakeholder communities.

The list of 34 global risks included in the survey was updated in 2023 as follows.

Seven new risks were added in response to observed trends across all five categories (economic, environmental, geopolitical, societal and technological):

1. Censorship and surveillance

2. Critical change to Earth systems (climate tipping points)

3. Concentration of strategic resources (minerals, materials)

4. Erosion of human rights and/or civic freedoms

5. Inequality or lack of economic opportunity

6. Intrastate violence (civil strikes, riots, coups)

7. Talent and/or labour shortages

In addition:

“Misinformation and disinformation” was recategorized from a societal to a technological risk, while “Disruptions to critical infrastructure” was recategorized from a technological to an economic risk.

“Adverse outcomes of AI technologies” was split out from “Adverse outcomes from frontier technologies (quantum, biotech, geoengineering)”, while “Natural disasters and extreme weather events” was delineated into two separate categories (“Extreme weather events” and “Non-weather-related natural disasters”).

Global risk categories relating to a failure in governance were removed. This includes “Ineffectiveness of multilateral institutions and international cooperation”, as well as “Failure of climate-change adaptation” and “Failure to mitigate climate change”. These updates were made to ensure the global risks list focused on the core risk itself, rather than a related exposure or vulnerability to that risk through in/action.

The names and definitions of the remaining risks have been revised and, where applicable, merged, modified and/or expanded to reflect new ways in which the risks may materialize and the potential adverse outcomes they may cause. However, to ensure comparability over time, the fundamental concept of each risk has remained broadly consistent with that of previous versions of the survey, although names and definitions were modified.

The GRPS 2023–2024 was further refined this year to gather more granular perceptions of risk and to incorporate new approaches to risk management and analysis. To that end, the GRPS 2023–2024 was comprised of seven sections:

– Current risk landscape asked respondents to select up to five risks among 20 pre-selected risks that they believe are the most likely to present a material crisis on a global scale in 2024. The final rank is based on the share of respondents who selected the particular risk. The 20 options included: Accidental or intentional nuclear event; Accidental or intentional release of biological agents; AI-generated misinformation and disinformation; Attacks on critical infrastructure; Censorship and erosion of free speech; Cost of living crisis; Cyberattacks; Disrupted supply chains for critical goods or resources; Disrupted supply chains for energy; Disrupted supply chains for food; Economic downturn; Escalation or outbreak of interstate armed conflict(s); Extreme weather events; Housing-bubble burst; Institutional collapse within the financial sector; Public-debt distress; Skills or labour shortages; Societal and/or political polarization; Tech bubble burst; and Violent civil strikes and riots. Respondents were also able to write in additional risks to Other, a free-text field. Results are illustrated in Figure 1.2.

– Short- and long-term risks landscape asked respondents to estimate the likely impact (severity) of each of the 34 global risks, on a 1-7 scale [1 – Low severity, 7 – High severity], over both two-year and 10-year periods. “Severity” is meant to take into consideration the impact on populations, the economy or environmental resources on a global scale. Respondents were also allowed to nominate any other risk considered missing from the 34 global risks. A simple average based on the scores selected was calculated and the results are illustrated in Figures 1.3 and 2.2. In addition, if a respondent selected the highest severity score (7) for any of the 34 risks, they were asked a follow-up question to identify areas of particular concern with respect to the identified risk.

– Consequences seeks to understand the potential consequences of risks, to create a network map of the global risk landscape. Respondents were provided 10 randomly selected global risks (from the full list of 34 global risks) and were then asked to select up to five global risks (from the full list) likely to be triggered by each of the 10 randomly selected risks. Results are illustrated in Figure 1.7. In visual results, “Nodes: Risk influence” is based on a simple tally of all bidirectional relationships identified by respondents. “Edges: Relative influence” is based on a simple tally of the number of times the risk was identified as a consequence. However, visual results do not show all connections: weaker relationships identified by less than 25% of respondents were not included as edges.

– Risk governance asked respondents to identify approach(es) that they expect to have the most potential for driving action on risk reduction and preparedness over the next 10 years, with respect to the most severe risks (severity score of 6 or 7 over the 10-year timeframe). Respondents could choose up to three answers from the following nine approaches: Financial instruments (e.g. insurance, catastrophe bonds, public risk pools); National and local regulations (e.g. environmental, operational, financial regulations and incentives); Minilateral treaties and agreements (e.g. Basel, Wassenaar, regional free trade agreements); Global treaties and agreements (e.g. UNFCC, Paris, Montreal, NPT, WTO); Development assistance (e.g. international aid for disaster risk response and reduction); Corporate strategies (e.g. ESG reporting, resilient supply chains, social initiatives, PPPs); Research and development (e.g. new technologies, early-warning systems, global risk research); Public awareness and education (e.g. campaigns, school curricula, media products); Multistakeholder engagement (e.g. platforms for exchanging knowledge, best practices, alignment). A simple tally of the number of times an approach was identified was calculated for each risk. Results are illustrated in Figure 3.5. To ensure legibility, the names of some of the global risks have been abbreviated in the figures. The portion of the full name used in the abbreviation is in bold.

– Risk outlook asked respondents to characterize the evolution of the global risks landscape based on a number of factors. It first asked respondents to indicate which statement best characterizes current and future global efforts to manage the Earth’s resources. Respondents were provided with the same 7-point Likert scale for both the current and future timescales, ranging from “We need to respect Earth’s limits and restrict the consumption of natural resources to make our lifestyles sustainable” (1) to “We need to change Earth’s limits using science and technology to increase the supply of goods to create abundance” (7). A simple tally for each of the seven options was calculated.

– Respondents were then asked to select a statement that they believe best characterizes the global political environment for cooperation on global risks in 10 years. Respondents were provided with four options: (1) Continuation or reinvigoration of the US-led, rules-based international order; (2) Multipolar or fragmented order in which middle and great powers contest, set and enforce regional rules and norms; (3) Bipolar or bifurcated order shaped by strategic competition between two superpowers; (4) Realignment towards a new international order led by an alternative superpower. A simple tally for each of the four options was calculated. Results are illustrated in Figure 2.31.

– Finally, respondents were asked to select a statement which best characterizes their outlook for the world over the next two and 10 years. Respondents were provided with the same five options for both time periods: (1) Calm: negligible risk of global catastrophes; (2) Stable: isolated disruptions, low risk of global catastrophes; (3) Unsettled: some instability, moderate risk of global catastrophes; (4) Turbulent: upheavals and elevated risk of global catastrophes; (5) Stormy: global catastrophic risks looming. A simple tally for each of the five options was calculated. Results are illustrated in Figure 1.1.

A total of 1,852 responses to the GRPS were received. From these, 1,490 were kept, based on the threshold at least one non-demographic answer, a minimum answer time of 2 minutes, and the filtering of multiple submissions based on browser cookies as well as partial responses (>40%) that have overlapping IP-numbers and demographic answers with a fully recorded response (100%).

– Current risk landscape: 1,490 respondents selected at least one risk.

– Short- and long-term risks landscape: 1,312 respondents evaluated the severity of at least one risk in one time frame.

o Short-term severity: 1,312 respondents evaluated the severity of at least one risk.

o Long-term severity: 1,311 respondents evaluated the severity of at least one risk.

– Consequences: 1,049 respondents paired at least one risk with one consequence.

– Risk governance: 952 respondents selected at least one approach for at least one risk.

– Risk outlook: 1,001 respondents answered at least one question.

o Global efforts: 984 respondents answered over at least one time frame.

o Global political environment for cooperation: 981 respondents answered.

o Outlook for the world: 992 respondents answered over at least one timeframe.

– Sample distribution: 1,490 respondents who answered at least one non-demographic question were used to calculate the sample distribution by place of residence (region), gender, age, area of expertise and organization type.

Figure B.1 presents some key descriptive statistics and information about the profiles of the respondents.