By 2100, how will the percentage of people who believe in God evolve?

This question explores the change in belief in God over time, focusing on whether there will be a noticeable increase in the percentage of people who report believing in a deity by the end of the century.

Forecasted on

Forecast probabilities

  • Significant increase: 18.8%
  • No change: 31.3%
  • Significant decrease: 50%

Reasoning

A significant decrease in the global share of people who believe in God looks most likely by 2100, but not by a landslide. The basic reason is simple: modernization, education, and generational replacement have a long history of wearing down belief, yet demography and instability keep religion from fading nearly as fast as headlines from rich countries suggest. That leaves no change as a substantial second possibility, and a significant increase as plausible but clearly the least likely.

Probability judgment: Significant decrease 50% | No change 31% | Significant increase 19%

Significant decrease: modernization is a slow, relentless solvent

If you wanted to build the case for a lower global belief rate by 2100, you would start with the most durable finding in the sociology of religion: as societies get richer, more educated, more urban, and more individualistic, belief tends to weaken. That doesn’t happen everywhere at the same speed, and it doesn’t happen in a straight line. But over three-quarters of a century, those forces compound.

The recent evidence points in that direction. South Korea offers the cleanest snapshot of where hyper-modern societies can end up: three out of four people in their 20s identifying as having no religion. Reports on younger women “breaking free” from religion, and on the weakening transmission of belief from parents to children, tell a similar story. Even some supposed signs of revival have looked thin on inspection, as when the British Bible Society had to retract a claim about a quiet resurgence among young adults [e103, e169, e56, e204].

Education matters here, not because schooling automatically makes people unbelievers, but because it broadens the set of explanations people use to make sense of the world. The background material points to expanding global education and the spread of more scientific, secular frameworks. And now AI is becoming part of that process: the AI-in-education market is expected to grow rapidly, AI is moving from analytical support to autonomous discovery in science, and adoption is broadening well beyond the tech elite. That is exactly the kind of cognitive environment in which traditional religious authority often loses some of its monopoly on meaning [e87, e231, e243, e257].

Technology also changes what religion used to be especially good at providing: explanations, moral guidance, even a sense of presence and intimacy. There was a piece asking whether AI has made God unnecessary, another on “generative charisma” in religious practice, and still others describing AI companions, AI Jesus, and a devotional metaverse built around Sanatan Dharma. Those developments don’t eliminate belief. But they do suggest that the institutions and habits through which belief has historically been maintained are entering far more competitive territory [e2, e1, e45, e48, e101].

Then there is the generational arithmetic. Young people are getting information from very different sources than their parents, and cultural secularism often advances less through dramatic conversions than through quiet non-transmission. Once belief stops being socially assumed, each generation starts from a lower baseline than the last. Over a horizon as long as 2100, that matters enormously [e78, e56].

A final reason to take decline seriously is that many of today’s most religious regions are also modernizing. Africa now accounts for 12 of the world’s 20 fastest-growing economies, African banks are outperforming other regions in profitability, Morocco has become the continent’s leading industrial economy, and national AI strategies are spreading across Africa’s main tech hubs. India, meanwhile, is still posting strong growth even amid geopolitical stress. The point is not that growth instantly secularizes societies. It’s that the parts of the world that currently anchor global belief are not standing still, and over decades modernization can chip away even where starting religiosity is very high [e232, e245, e253, e256, e251].

That is why significant decrease ends up as the most likely outcome.

No change: the demographics of belief are stubborn

If the secularization story were the whole story, this forecast would be easier. It isn’t, because population growth is not evenly distributed, and religion has a demographic advantage.

The most important counterweight is that future population growth is expected to be concentrated in places that remain highly religious, especially Sub-Saharan Africa and parts of South Asia. Meanwhile, many of the world’s most secular societies are also the ones with the lowest fertility and, in some cases, outright population decline. Greece has been shrinking for 15 straight years. New Zealand’s fertility has hit a record low. South Korea remains the poster child for demographic collapse. That means even if rich societies continue secularizing, their share of humanity will likely fall [e57, e136, e5, e107, e4].

This is the arithmetic that makes “no change” so plausible. Global belief is not the same as OECD belief. The world average can stay roughly flat even while Europe and East Asia keep drifting away from religion, simply because the future will include more Africans and South Asians than today.

There is also a measurement issue that favors stability. “No religion” is not the same thing as “no belief in God.” Institutional affiliation often falls first; belief lingers longer. The category in this question is broad, and broad measures of belief are usually stickier than church, mosque, or temple attendance. That pushes probability away from a dramatic collapse.

Just as important, religion has a habit of thriving when life feels precarious. And the news flow offers plenty of precariousness: Islamist militants gaining strength in parts of Africa, a U.S.–Iran war reshaping communications and supply chains, fertilizer prices rising because of shipping disruptions, slower global growth forecasts, and harsher droughts, floods, heatwaves, and pressure on food and water supplies. Whatever one thinks of the secularization thesis, it has always had an asterisk: prosperity can reduce religiosity, but insecurity often sustains it [e233, e248, e238, e229, e250, e55].

Religion also appears more adaptable than some secularization narratives assume. Tech companies have been turning to faith leaders for advice on AI ethics. OpenAI and Anthropic met with religious leaders. Articles in the Associated Press and elsewhere described Silicon Valley looking to religion as a source of moral guidance. A cardinal in Krakow argued that AI needs a moral compass, and even said faith is growing in secular Europe. This isn’t proof of revival, but it is evidence that religion can reposition itself when the social environment changes instead of simply retreating [e11, e15, e16, e19, e177].

The same goes for politics. Governments and movements continue to use religion as a source of legitimacy, identity, and mobilization. There was reporting on public-awareness campaigns aimed at strengthening ties between governments and religious organizations, on anti-Islam rhetoric framed as a battle between Jesus Christ and Satan, on states invoking religion during the war in Iran, and on authoritarian pressure on religious communities in China. None of that guarantees sincere belief. But it does keep religious language and identity at the center of public life, which helps explain why religion often proves more persistent than tidy modernization models predict [e230, e234, e199, e105, e117].

All of that makes no change a very live possibility: secularization in some places, demographic reinforcement in others, and enough crisis and adaptation to keep the global average from moving much.

Significant increase: possible, but the ceiling is low

The argument for a meaningful increase in the share of believers is not crazy. In fact, it has one very strong pillar: the world’s demographic center of gravity is moving toward more religious societies.

If Sub-Saharan Africa and parts of South Asia retain very high levels of belief while becoming a much larger share of the world’s population, the global percentage could rise. Extremely low fertility among secular populations strengthens that mechanism. Add in ongoing conflict, climate stress, migration tensions, identity politics, and technological anxiety, and you have the ingredients for religion to remain not just relevant, but more salient. There are scattered signs of renewed interest in religion in the United States, commentary about Gen Z reshaping spirituality, and claims of growth in parts of Europe [e57, e104, e250, e12, e200, e177].

There is also a subtler version of the bullish case: not that old-style religion comes roaring back, but that new forms of spiritual practice make belief more accessible and personalized. The AI devotional metaverse, AI-mediated religious experiences, and the broader fusion of tech and spirituality could, in theory, lower barriers to engagement rather than raise them [e101, e1, e45].

Still, this is the least likely of the three outcomes for a pretty basic reason: belief already starts from a high level in much of the developing world. That creates a ceiling effect. A region where 90-plus percent already believe in God cannot contribute much to a global increase, but it has plenty of room to decline. Over a 74-year horizon, even modest secularization in today’s high-belief societies could cancel out the demographic gains from their population growth.

And the supposed revival stories have often looked weaker under scrutiny than they first appeared. The retracted British “revival” claim is a useful cautionary tale. So are the sharp secular trends among youth in places like South Korea. It is easier to imagine religion proving resilient than to imagine a broad global surge in belief large enough to count as a significant increase [e204, e103].

That is why significant increase stays below one chance in five.

Why the balance lands at 50/31/19

The cleanest way to think about this forecast is that two huge machines are running at once.

One machine is modernization: more education, more urbanization, more individualism, more technological mediation of daily life, and more generations raised in cultures where religion is optional rather than assumed. That machine points toward lower belief, and over a period stretching to 2100 it has a lot of time to do its work [e87, e231, e243, e56].

The other machine is demographic and political resilience: faster population growth in religious regions, lower fertility in secular ones, persistent insecurity, and the continuing usefulness of religion as a source of identity, community, and legitimacy. That machine slows decline dramatically and makes outright stability plausible [e57, e136, e4, e5, e233, e230].

On balance, the first machine looks a bit stronger by 2100, especially because the regions now driving global population growth are also beginning to experience the very forces that have eroded belief elsewhere. But the second machine is strong enough that a flat global share remains a substantial possibility. What it does not do, at least on the evidence here, is make a clear global increase the likeliest outcome.

So the best synthesis is: - Significant decrease: 50% - No change: 31% - Significant increase: 19%

Cited evidence