Has AI become the world’s fastest growing research field?

---
title: "Has AI become the world's fastest growing research field?"
output:
  flexdashboard::flex_dashboard:
    orientation: rows
    vertical_layout: scroll
    social: menu
    source_code: embed
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = FALSE, message = FALSE, warning = FALSE)
source("build_dashboard.R")
library(htmltools)
```

<style>
body {
  font-family: Arial, sans-serif;
  color: #1A1A1A;
  background: #f5f5f5;
}

.navbar {
  background-color: #1A1A1A !important;
  border-bottom: 3px solid #D9281A;
}
.navbar-brand {
  color: #FFFFFF !important;
  font-family: Georgia, serif;
  font-weight: bold;
}
.navbar-inverse .navbar-nav > li > a {
  color: #CCCCCC !important;
}

.story-intro {
  max-width: 600px;
  margin: 0 auto;
  padding: 24px 16px 16px;
}
.story-intro h1 {
  font-family: Georgia, serif;
  font-size: 1.85rem;
  line-height: 1.25;
  color: #1A1A1A;
  margin-bottom: 6px;
  border-left: 4px solid #D9281A;
  padding-left: 12px;
}
.story-intro .subhead {
  font-size: 1rem;
  color: #555555;
  font-style: italic;
  margin-bottom: 16px;
  padding-left: 16px;
}
.story-intro p {
  font-size: 0.97rem;
  line-height: 1.65;
  color: #333333;
}

.chart-caption {
  max-width: 600px;
  margin: 4px auto 12px;
  padding: 6px 10px;
  font-size: 0.82rem;
  color: #666666;
  border-top: 1px solid #E0E0E0;
  line-height: 1.5;
}
.chart-caption em {
  font-style: normal;
  color: #888888;
}

.plot-box {
  max-width: 600px;
  margin: 0 auto;
}

.chart-title {
  font-weight: 600;
  color: #1A1A1A;
}

.story-conclusion {
  max-width: 600px;
  margin: 0 auto;
  padding: 20px 16px;
}
.story-conclusion h2 {
  font-family: Georgia, serif;
  font-size: 1.4rem;
  color: #D9281A;
  border-bottom: 2px solid #F2F2F2;
  padding-bottom: 8px;
  margin-bottom: 14px;
}
.story-conclusion p {
  font-size: 0.96rem;
  line-height: 1.65;
  color: #333333;
}

.references {
  max-width: 600px;
  margin: 0 auto;
  padding: 16px;
  background: #F9F9F9;
  border-left: 3px solid #D9281A;
  border-radius: 2px;
}
.references h3 {
  font-family: Georgia, serif;
  font-size: 1.1rem;
  color: #1A1A1A;
  margin-bottom: 10px;
}
.references ul {
  padding-left: 18px;
  margin-bottom: 12px;
}
.references li {
  font-size: 0.88rem;
  color: #444444;
  margin-bottom: 6px;
  line-height: 1.5;
}
.ai-acknowledge {
  font-size: 0.85rem;
  color: #666666;
  border-top: 1px solid #E0E0E0;
  padding-top: 10px;
  margin-top: 12px;
}
</style>


Row
-------------------------------------

### Introduction {.no-title}

<div class="story-intro">
<h1>Has AI become the world's fastest growing research field?</h1>
<p class="subhead">What the open scientific record tells us about the global AI boom</p>

<p>
Spend five minutes on any university campus and you'll hear it — AI is everywhere. Students are
using it, lecturers are debating it, and research labs are racing to publish on it. But how big
has this actually gotten? And who is really driving it?
</p>

<p>
This story digs into the open scientific record to find out. Using live data from
<a href="https://api.openalex.org" target="_blank">OpenAlex</a> (OpenAlex, 2026) and the
<a href="https://data.worldbank.org/indicator/NY.GDP.PCAP.CD" target="_blank">World Bank</a>
(World Bank, 2026), the five charts below trace how AI research has grown since 2000, which
countries and universities lead it, which topics are taking off, and whether money explains who
gets to be at the frontier.
</p>
</div>

Row
-------------------------------------

### Chart 1 — Growth Over Time

```{r chart1, fig.width=6, fig.height=4.6}
p1
```

Row
-------------------------------------

### {.no-title}

```{r caption1}
HTML('
<div class="chart-caption">
The chart displays two filled area lines tracking AI publication volume from 2000 to 2025 —
one for journal and review articles, one for conference papers and preprints. Both series sit
close together through the early 2000s, then begin pulling apart around 2015 as the deep
learning wave gathered pace. The steepest climb comes after 2022, when the public release of
large language models triggered a surge of research activity across nearly every discipline.
By 2025, conference papers and preprints are outpacing journals, reflecting how fast the field
now moves — formal peer review simply cannot keep up with the rate of new work being produced.
<br><br>
<em>Source: OpenAlex Works API, filtered by concept: artificial intelligence (OpenAlex, 2026).</em>
</div>
')
```

Row
-------------------------------------

### Chart 2 — Which Countries Lead?

```{r chart2, fig.width=6, fig.height=4.6}
p2
```

Row
-------------------------------------

### {.no-title}

```{r caption2}
HTML('
<div class="chart-caption">
This horizontal bar chart ranks the top 15 countries by total AI publication count, ordered
from lowest to highest so the largest producers sit at the top. The United States and China
stand far ahead of everyone else — their bars are roughly double the length of third-place
India. Europe is well represented, with Germany, France, the UK, Italy and Spain all making
the list, but none individually come close to matching the US-China gap. The pattern points to
a research landscape where two countries set the agenda, and the rest of the world responds
to it rather than driving it.
<br><br>
<em>Source: OpenAlex Works API, grouped by author country (OpenAlex, 2026).</em>
</div>
')
```

Row
-------------------------------------

### Chart 3 — Which Institutions Publish the Most?

```{r chart3, fig.width=6, fig.height=4.6}
p3
```

Row
-------------------------------------

### {.no-title}

```{r caption3}
HTML('
<div class="chart-caption">
The treemap fills its area with 15 rectangles, each sized in proportion to that institution\'s
total AI publication count. The Chinese Academy of Sciences takes up a notably large tile,
as does MIT, Stanford and a cluster of UK universities. What stands out is how many of the
biggest tiles belong to state-backed research bodies rather than traditional universities —
CNRS in France, for instance, sits alongside elite private institutions. Hover over any cell
to see the exact figure. The concentration visible here mirrors what Chart 2 shows at the
country level: a small number of organisations are responsible for a disproportionate share
of what the world knows about AI.
<br><br>
<em>Source: OpenAlex Works API, grouped by author institution (OpenAlex, 2026).</em>
</div>
')
```

Row
-------------------------------------

### Chart 4 — Which Topics Are Taking Off?

```{r chart4, fig.width=6, fig.height=4.6}
p4
```

Row
-------------------------------------

### {.no-title}

```{r caption4}
HTML('
<div class="chart-caption">
Six topic lines run from 2015 to 2025, each tracking a different AI subfield. Machine learning
and deep learning start highest and grow steadily, but their curves are relatively smooth
compared to what happens with generative AI from 2022 onward — that line turns almost vertical,
dwarfing everything else by 2025. Natural language processing shows a similar but smaller jump
around the same period, driven by the same wave of large language model research. Robotics, by
contrast, grows at a modest and consistent pace throughout. Click any label in the legend to
isolate or hide individual topics, which makes the generative AI spike even more striking when
viewed on its own.
<br><br>
<em>Source: OpenAlex Works API, publication counts by topic and year (OpenAlex, 2026).</em>
</div>
')
```

Row
-------------------------------------

### Chart 5 — Does Money Predict Output?

```{r chart5, fig.width=6, fig.height=4.6}
p5
```

Row
-------------------------------------

### {.no-title}

```{r caption5}
HTML('
<div class="chart-caption">
The scatter plot places each of the top 15 countries on a log-log scale, with GDP per capita
on the x-axis and AI publication count on the y-axis. Bubble size encodes population, so
large countries like China, India and the United States are immediately visible. A loose
upward trend is present — wealthier nations do tend to publish more — but the outliers tell
the more interesting story. China and India both sit well to the upper-left of where the
trend line would predict, showing that sheer scale of investment and population can compensate
for lower income per head. Smaller high-income countries like the Netherlands cluster toward
the right but with far fewer publications, suggesting that wealth alone is not enough without
the institutional infrastructure to back it up.
<br><br>
<em>Sources: OpenAlex Works API (OpenAlex, 2026); World Bank indicator NY.GDP.PCAP.CD
(World Bank, 2026).</em>
</div>
')
```

Row
-------------------------------------

### Conclusion {.no-title}

<div class="story-conclusion">
<h2>The AI boom is real — but it isn't evenly shared</h2>

<p>
Taken together, these five charts tell a pretty clear story. AI research has grown faster than
almost any field in the open scientific record, and the pace has picked up even more since
generative AI went mainstream around 2022. But this growth is not spread evenly. A small group
of countries, and within those countries a small group of universities and research agencies,
are producing the vast majority of what gets published (OpenAlex, 2026).
</p>

<p>
For students today, that matters more than it might seem. The research that gets done shapes
the tools that get built, which shapes the jobs, the policies and the everyday experiences that
follow. If the people deciding what AI research to pursue are concentrated in a handful of
places, it's worth asking whose needs and values are getting built into the systems the rest of
the world ends up using.
</p>

<p>
The economic picture adds a final nuance: wealth clearly helps, but it isn't destiny. China and
India have shown that deliberate investment in research infrastructure can close the gap — even
at lower income levels (World Bank, 2026). The question is whether other countries will follow.
</p>
</div>

Row
-------------------------------------

### Data and References {.no-title}

<div class="references">
<h3>Data and References</h3>

<p>
All five charts pull live data from open APIs when the dashboard renders. Charts 1–4 query the
OpenAlex Works API by concept, grouping results by year or institution. Chart 5 combines those
country-level counts with GDP per capita and population from the World Bank. If an individual
API call fails at render time, that value falls back to a June 2026 snapshot so the charts
always display. Because the data updates continuously, exact figures may differ slightly between
render dates.
</p>

<ul>
  <li>
    OpenAlex. (2026). <em>Works API</em>. OurResearch.
    <a href="https://api.openalex.org/works" target="_blank">https://api.openalex.org/works</a>
  </li>
  <li>
    World Bank. (2026). <em>World Development Indicators: GDP per capita, current US$
    (NY.GDP.PCAP.CD)</em>. World Bank Group.
    <a href="https://data.worldbank.org/indicator/NY.GDP.PCAP.CD" target="_blank">https://data.worldbank.org/indicator/NY.GDP.PCAP.CD</a>
  </li>
  <li>
    World Bank. (2026). <em>World Development Indicators: Population, total (SP.POP.TOTL)</em>.
    World Bank Group.
    <a href="https://data.worldbank.org/indicator/SP.POP.TOTL" target="_blank">https://data.worldbank.org/indicator/SP.POP.TOTL</a>
  </li>
</ul>
</div>