AI Agents Are About To Change Gaming Forever

Video game developers have always pushed technology to its limits to build worlds that feel alive, and AI is already reshaping how those worlds get made.

Now, a bigger leap is on the horizon: AI agents that can plan, act, and carry out multi-step tasks, meaning game characters can start behaving less like scripted props and more like believable participants.

That opens the door to richer interactions, more unpredictable gameplay, and lower barriers for smaller studios. It also sharpens the risks, from agents being manipulated into toxic behavior, to subtle pressure toward in-game spending, to fresh anxieties about creative jobs and who really controls the story.

So what does the next era of interactive entertainment look like when agents are built into the game world itself, and what do developers need to get right before this becomes mainstream?

How Are Agents Being Used in Gaming?

So far, the agent integration in mainstream gaming has focused on augmenting the ability to create lifelike or interesting non-player characters (NPCs).

Algorithms controlling the actions of NPCs and enemies in video games have always been referred to as “AI”, but in fact, they usually follow scripts and react to the player’s actions, usually with some randomness thrown in.

With the arrival of agentic AI, however, characters can offer near-human levels of interaction and engagement.

For example, Fortnite recently incorporated an agentic Darth Vader character into its phenomenally successful multiplayer game world. The character can converse with players, as well as choose to either join them to help them accomplish their goals or fight against them.

Sony has introduced Sophy, its "superhuman AI racing agent," into the world of Gran Turismo, which can talk and communicate with real players as they compete against her in online league tables.

Another gaming use case for AI agents that’s expected to prove popular is as referees and game masters. This could involve using an AI to take oversight of a game, with the aim of ensuring it’s fair, interesting and entertaining.

What’s most interesting is that this agentic behavior is likely to lead to “emergent” forms of gameplay. This is where gamers find ways to engage and entertain themselves in games that the developers may not have ever thought about.

Imagine a living-world game like Grand Theft Auto where every character you encounter on the street has their own goals and motivations, and can converse and interact just like a real person. The possibilities are staggering.

This All Sounds Great. What Could Go Wrong?

Well, it’s fair to say that the conflux of AI and gaming has stirred some controversies. And with agentic AI crossing yet another line by actually carrying out tasks traditionally reserved for humans, there could be even more trouble ahead.

For a start, it’s often apparent that the technology isn’t “quite there” yet. AI hallucination remains a very real problem, and agents, despite their superior reasoning and planning abilities, remain highly vulnerable to it.

Their behavior is also often vulnerable to manipulation. The Fortnite Darth Vader mentioned in the previous section, for example, was quickly subverted to get it to swear and use homophobic slurs.

And there are also concerns that this technology could be used to manipulate us into in-game spending, perhaps by befriending us or even threatening to prevent us from winning.

Agentic Vader is controversial in other ways, too, sparking objections from groups representing human actors due to its use of a synthetic AI voice.

This hints at an even darker side (excuse the pun) of agentic AI in gaming: its potential impact on human jobs.

Agentic AIs are frequently referred to as “virtual workers”. In an industry that has recently seen wide-scale redundancy, there is a great deal of sensitivity around the subject of global corporations like Microsoft, Sony and Electronic Arts using AI to do work that could provide a living for humans.

On the other hand, with agentic AI tools like Manus making it easier to create games without spending hours writing code, agents are seen by many as dramatically lowering the barrier when it comes to game development.

In theory, it means that smaller studios or even solo developers can generate code bases and in-game assets like graphics and sounds in a fraction of the time it would have taken before.

This could level the playing field when competing with far larger multinational developers producing AAA titles and could lead to a new wave of creativity and innovation.

The Future Of Interactive Entertainment

Looking ahead, it’s clear that the collision of agentic AI and video gaming could be massive.

The ability to converse and interact with characters that act just like us will open up new frontiers for interactive experiences. But there are challenges that need to be addressed, too.

The behavior of AI agents still isn’t fully understood, so before we let them loose in games, should we take steps to minimize the risk that they might attempt to manipulate us? It seems unethical, for example, to allow agents to appear to befriend us, or even form a meaningful attachment, in order to persuade us to spend money.

And does trusting them to “tell the story” by dictating the direction that gameplay will take through their actions and decisions equate to delegating creative control?

As thousands of games industry creatives who feel their craft and livelihood are at risk have made clear, the drive for innovation should always be balanced with respect for human creativity.

If these questions can be answered, then agentic AI has the potential to transform gaming forever, giving us the ability to explore infinitely detailed worlds, peopled by deeply realistic, simulated people. Something not too dissimilar to the Star Trek holodeck, maybe.

Getting it wrong, on the other hand, could lead to misery as creative jobs are vacuumed up by digital workers and a wave of AI-generated entertainment that many humans will find difficult to connect with emotionally.