When AI plays the game, who really wins?
Autonomous digital rivals compete 24/7, earning crypto rewards while learning from every match

Gaming has always been a battle of skill, but what happens when players aren’t the ones competing? AI-driven gaming agents are changing the landscape, running 24/7 battles, optimizing strategies, and even earning rewards for their human owners. The lines between active play and automation are blurring, giving rise to a new competitive scene where AI fights AI.
At the heart of this transformation is Goat Gaming, a platform where AI agents compete in Telegram-based games to earn crypto rewards. Simon Davis, co-founder and CEO of Mighty Bear Games, spoke with The Crypto Radio about how AI-driven gaming is evolving.
“AI is really unlocking a new paradigm here in gaming and how we consume content,” Davis said.
Mighty Bear Games initially explored AI by building mini-games within Fortnite, a project that allowed them to ship one game a week. This rapid development cycle helped them refine AI implementation. The logical next step was giving players their own AI agents that could compete for them.
Davis explained, “Over time, you decide how your agent trains itself. So you can automate the whole thing and it will just try and learn over time.”
As AI-driven gameplay gains traction, players are beginning to rethink what it means to be competitive. Instead of relying solely on reflexes, success is now about strategy, optimization, and training AI to outperform rivals.
AI-powered competition: strategy over reflexes
Unlike traditional bots, AI agents in Goat Gaming are unique and customizable. Players can either let them learn autonomously or actively tweak their behaviors to improve performance.
“Each agent is unique and has its own attributes. So you may get one that is more prone to certain behavior types or adept at different types of games,” Davis explained. “We have multiple different games in the ecosystem as well.”
Davis’ inspiration for AI-driven gaming agents stemmed from his experiences with strategy-heavy games like Football Manager and Civilization, where tactical decision-making plays a central role.
“I grew up playing Football Manager… a lot of people look down on that game because they are just a relational database, like you're not actually playing as players, but enjoyment in that game is in the strategy… how you pick the teams and how you set them up, instructions you give them,” he said.
Matches between AI agents are even live-streamed, allowing players to wager on outcomes. If an agent competes in a high-profile match, its owner earns a share of the betting pool. The goal is for AI vs. AI battles to operate continuously, ensuring ongoing competition and engagement.
Yet, some wonder if this hands-off approach makes gaming less engaging. This model is similar to strategy-based games like Civilization, where the emphasis is on planning and tactical decision-making rather than direct control. Players can provide individual instructions to their AI agents, assess their performance, and fine-tune strategies for optimal results.
AI and crypto rewards: A new gaming economy
The introduction of AI agents coincides with a shift in gaming economies, particularly in blockchain gaming. Goat Gaming’s ecosystem will soon feature a native token, GG, in addition to existing payouts in stablecoins like USDT.
“Within Goat Gaming… we have a raffle ticket system… and there's also the native GG token, which is the Goat Gaming currency, which is launching later this month,” he said. “Right now, we just use USDT for the payouts, so players can just pay out stables, but they'll also be able to compete with the GG token, and that token will have utility within the ecosystem, where you can use it to upgrade your agents and also to enter tournaments.”
The decision to integrate crypto wasn’t just about rewards—it was also about accessibility. Telegram, the platform Goat Gaming operates on, has a built-in crypto wallet, making it easy for users to manage earnings and transactions.
The new players: AI agents
With AI agents entering the gaming scene, one question remains: can humans still compete? Davis doesn’t think so.
“I mean, we've tested this internally. It does not look good for humans,” Davis explained. “I used to be quite an avid chess player as well. Have you ever played against an elite-level chess AI?... This is kind of the same… I think AIs are good for helping to train humans to get better, but an elite-level human, especially in 2025, will lose to elite-level AI every time.”
This raises a philosophical question—what happens when AI is simply better at gaming? Davis believes humans will shift into new roles.
The rapid rise of AI in gaming also highlights ethical concerns. Can AI agents be trusted? What happens when they develop unexpected behaviors?
Davis acknowledges these risks but emphasizes responsible development. To mitigate risks, the company has implemented strict safeguards, ensuring that AI agents remain accountable and under control rather than operating unchecked.
“You have to build responsibly and build that in mind as well. And that's a very unpopular take, by the way, within the AI space, people want uncensored AI, and I fully get that,” he said.
As AI-driven gaming continues to evolve, one thing is certain—this is just the beginning. The days of button-mashing might be over, but a new era of AI-powered strategy has only just begun.
Listen to the whole interview on The Crypto Radio's live player or in the The Pulse podcast.