Embark Studios is set to redefine player-versus-enemy combat in its online shooter Arc Raiders with the arrival of the Bully, a new ground-based foe that learns and adapts through machine learning, arriving alongside the studio’s largest update to date, Frozen Trail. Scheduled for release on Thursday, October 8, the free update will launch simultaneously across PS5, Windows PC, and Xbox Series X, bringing a wave of new content and a fresh adversary that mirrors the behavior-learning mechanics already showcased by the game’s existing Leaper enemy. The move marks a notable step forward in how the developer approaches AI-driven opponents, blending research-grade machine learning with live-action gameplay to create more unpredictable and memorable encounters.
The concept of an enemy that improves through experience is not new to Arc Raiders, but the Bully represents the studio’s most ambitious application of the technology so far. During a virtual pre-briefing with press and content creators ahead of the Frozen Trail reveal, Polygon observed the new creature in action. Unlike the towering, long-legged Leaper, the Bully is stocky and compact, yet it carries a relentless aggression that defines its presence on the battlefield. According to the studio, the enemy will pursue players without pause, refusing to give them a moment to recover. The design intent is to generate the same kind of chaotic, dynamic moments that the Leaper already delivers—enemies stumbling, falling over, and crashing into players mid-fight.
A Legacy of Learning Enemies
The Bully builds directly on the foundation laid by the Leaper, one of the most talked-about enemies in the game’s roster. At GDC earlier this year, Embark Studios’ machine learning research lead Martin Singh-Blom explained that the Leaper learns its behavior through rewards and recognition, allowing it to act in conventional ways while also adapting to difficult situations such as losing a leg. While Singh-Blom noted that the technology is not quite the same as fully AI-led behavior, the similarities are unmistakable, and the Leaper’s long legs and movement capabilities have made it the most visually striking example of the studio’s approach.
Now, that same philosophy is being extended to a new archetype. The Bully is another ground-based enemy equipped with machine-learning abilities, though it is not as large as the Leaper. Its defining trait is persistence: rather than relying on sheer size or reach, it overwhelms players through constant pressure and the sheer unpredictability of its movements. The result is a foe that transforms tense firefights into chaotic scrambles, forcing players to react to an opponent that behaves in ways that cannot be perfectly predicted.
“We see machine learning as an emerging technology, which is still in its infancy to some degree,” said Aleksander Grøndal, executive producer of Arc Raiders. “It’s still a bit of an advanced research task on our end. The interesting thing about it is that you can’t control it super well. We try to teach it how to behave, and sometimes that doesn’t make sense at all. Sometimes we need to update the model to fix particular issues with it.”
Aleksander Grøndal, Executive Producer, Arc Raiders
Grøndal’s comments highlight the delicate balance Embark Studios is attempting to strike between emergent intelligence and controllable gameplay. The studio trains its models across multiple platforms, but the best results have come from enemies that possess legs they can manipulate as players—allowing the AI to learn how to walk, adapt to obstacles, and recover from setbacks like losing a limb or tumbling over. These scenarios are precisely what generate the “fun and dynamic moments” the team is chasing, turning routine encounters into something genuinely surprising.
The community has already taken notice of the studio’s willingness to experiment with machine learning, with many players expressing excitement over the prospect of facing opponents that genuinely learn from their tactics. The Bully’s relentless pursuit style has drawn particular interest, as it promises to test even the most seasoned raiders and reward those who can think on their feet rather than relying on predictable patterns.
Looking Ahead
Frozen Trail arrives as a free update, meaning existing players will receive the new content at no additional cost. The update lands on Thursday, October 8, for PS5, Windows PC, and Xbox Series X, ensuring all major platforms receive the same experience simultaneously. Beyond the Bully, the update is billed as the studio’s biggest yet, promising a substantial amount of new content to complement the introduction of this new machine-learning adversary.
Grøndal’s remarks also suggest that machine learning may not be a one-off feature. As the studio continues to refine its models and update them to address unexpected behaviors, players may see additional enemies adopt similar adaptive traits over time. For now, the Bully stands as a bold statement of intent from Embark Studios—one that positions Arc Raiders at the forefront of AI-driven combat in the live-service shooter space, and sets the stage for what could be a defining direction for the game’s future.
1 comment
[…] Article: Arc Raiders Unleashes Machine-Learning Bully in Biggest Update Yet, Promising to Turn Every Fight In… Kaynak: Kotaku […]