AI
Becomes Standard in Game Development
Generative AI is no longer an experiment in the gaming industry.
It has become embedded in development pipelines at studios of every size,
handling tasks that range from the unglamorous to the genuinely creative. A
new survey has put a specific number on the scale of adoption that many in
the industry had sensed but not yet quantified: 87 percent of game developers
are now using AI agents somewhere in their workflow.
The figure comes from a survey conducted by Google Cloud in
partnership with The Harris Poll, and reported by Reuters. The breakdown
reveals where the adoption is concentrated. Some 47 percent of developers use
AI for playtesting and game balancing, 45 percent for localisation of text
and dialogue into other languages, and 44 percent for scripting and code
generation. These are areas where AI’s ability to handle high volumes of
repetitive, structured work aligns closely with tasks that have historically
consumed significant developer time without contributing directly to the
creative distinctiveness of a title.
More striking is the figure for creative applications. More than a
third of surveyed developers are using AI for tasks including dialogue
writing, level design, and storytelling elements. This marks a meaningful
shift from the early framing of AI in game development as a tool for
operational efficiency. It is becoming a creative instrument as well, with
implications that the industry has only begun to work
through.
Dynamic NPCs and the Changing Player Experience
The area where generative AI is having the most visible effect on
players, as opposed to developers, is in non-player character behaviour.
AP News has documented
how studios are deploying AI to create NPCs capable of responding in real
time to player choices rather than cycling through pre-scripted dialogue
trees. An NPC in a traditionally designed game has a finite number of
responses; once you have encountered them all, the character becomes predictable
and the illusion of a living world erodes.
AI-driven NPCs change that fundamentally. Rather than selecting
from a library of scripted responses, they generate contextually appropriate
dialogue based on the player’s history within the game, the current state of
the world, and the character’s established personality. A merchant who has
watched you complete several quests for a rival faction will react
differently to your presence than they did at the game’s opening. That
responsiveness, previously achievable only through expensive hand-crafted
scripting, can now be produced at scale.
The accessibility implications extend beyond narrative immersion.
For players who find rigid dialogue systems cognitively demanding or who
benefit from more flexible communication styles, AI-driven characters can
adapt their language and pacing in ways that pre-scripted systems cannot.
This connects to a broader pattern in how AI is being applied to make digital
environments more responsive to the diversity of people who inhabit
them.
The Legal and Creative Ownership Questions
Widespread AI adoption in creative workflows does not arrive without
complications. When an AI system designs a level, writes dialogue for a
character, or generates a visual asset, questions of intellectual property
become genuinely difficult. A Harvard Journal of Entertainment and Sports Law
analysis has examined how existing IP frameworks struggle with the concept of
AI-generated creative work, particularly when the AI was trained on
copyrighted material and the output is commercially exploited by a studio
that did not produce the underlying training data.
These are not abstract legal questions. They have practical
consequences for studio contracts, for the rights of voice actors and writers
whose work may have been used to train dialogue systems, and for the
long-term relationship between large studios that can afford sophisticated AI
tools and the independent creators whose work has historically fed the
cultural ecosystem that games draw from. The
courts are still developing frameworks for how AI-generated content should be
treated legally, and the games industry will be shaped by those
frameworks as much as by the technology itself.
Economic Pressures Driving Adoption
It would be a mistake to frame AI adoption in game development
purely as a creative story. The economic pressures behind it are as
significant as the creative possibilities. Development budgets for AAA titles
now regularly exceed hundreds of millions of dollars, while the window for
commercial success has narrowed as more titles compete for player attention
across more platforms simultaneously. Studios face a structural tension
between the cost of production and the price players will accept for a
finished game.
AI addresses parts of that tension directly. Localisation for a
global release involves translating millions of words of dialogue, interface
text, and marketing copy. At current rates for professional translation, that
cost can reach seven figures for a major title. AI-assisted localisation does
not replace human translators entirely, but it can dramatically reduce the
volume of work that requires full professional attention, lowering costs
while maintaining quality on the elements that matter most.
The same logic applies to playtesting. Traditional playtesting
requires human testers to spend hundreds or thousands of hours exploring a
game’s content and reporting bugs. AI agents can perform the same coverage in
a fraction of the time, freeing human testers to focus on the qualitative
judgements that require genuine player experience rather than systematic
coverage. The efficiency gains are real, even before the creative
applications of generative AI are factored in.
What the 36 Percent Using AI for Creative Tasks
Reveals
The figure that warrants the most attention in the survey data is
the 36 percent of developers using AI for creative storytelling tasks and the
37 percent using it to experiment with gameplay mechanics. These are not
peripheral applications. They are at the core of what makes a game distinctive.
Studios using AI for narrative design are not simply automating
existing processes. They are exploring new creative possibilities that were
previously impractical. Procedurally generated story branches, adaptive
difficulty curves that respond to individual player behaviour, and emergent
narrative systems that produce different stories for different players all
become more achievable when AI can generate and evaluate content at scale.
Just
as nations are discovering that AI can be a tool of influence at the
geopolitical scale, game developers are discovering that it can be
a tool of creative scale at the level of a single title.
The risk is that scale without quality produces experiences that
feel abundant but hollow. Players notice when dialogue lacks the specificity
and coherence of human writing, even if they cannot always articulate why.
The
psychological dimension of how people relate to AI-generated
content is becoming a serious area of research precisely because
the gap between what AI can produce and what humans find genuinely engaging
is not as wide as enthusiasts hoped but not as narrow as the technology’s
critics feared. For game developers navigating that gap, the survey’s 87
percent adoption figure is less a moment of arrival than the starting line of
a longer and more complicated creative journey. The tools are now widely
available; the harder work of learning what to build with them, and what to
protect from them, is only beginning.
About the Author
By Stuart Kerr, Technology Correspondent, LiveAIWire. Stuart
covers artificial intelligence, technology strategy, and the creative
industries being reshaped by machine learning. About
LiveAIWire.