Co-Workers, Co-Pilots, and Evolving Complexity for Game Developers

Game development is an intricate process to create an experience that blends complexity with intuitiveness, striking a balance that captivates both beginner and skilled players alike. With new innovations around Artificial Intelligence (AI), we are entering a new phase of game development that will include features such as AI Agents.

Across the NPC (non-playable character) landscape, there is a spectrum of complexity from simple "bots" to complex and sophisticated "agents" that support gameplay. An AI Agent refers to a more advanced, generative AI-powered character that can engage in dynamic, contextual conversations, and even show awareness of its surroundings. These agents can forge relationships with players and add depth and different dynamics to gameplay.

Agents in Games

While AI agents can add new dimensions to gameplay, their inclusion introduces new challenges to developers, who are tasked with balancing the creation of an engaging experience while not taking away the player’s rewarding journey of mastery. Today, agents can be broken into two separate buckets: Co-Workers and Co-Pilots.

In the context of AI agents for games, "co-pilot" and "co-worker" refer to roles that these agents can play while interacting with real players or other AI agents, each fulfilling distinct functions and having different interaction dynamics within the game environment. These distinct functions are core to the underlying gameplay and can have dramatic effects on gamer retention over time.

The inclusion of an AI agent (whether a co-worker or co-pilot, defined below) can lead to unplanned and unpredictable game experiences that the original game developer did not have in mind (this is a good thing). This allows the player to create unique, new, and innovative experiences in real time as they navigate the game’s objectives, missions, or narratives. This shift in core gameplay is significant because it allows for variety and a broadened experience for the gamer beyond the original game’s intent, which historically has primarily been a linear experience.

Co-pilot AI Agent

• Role and Function: A co-pilot AI agent is designed to assist the player directly, acting as a support system or advisor. This agent is not as focused on participating in core gameplay loops. This agent can provide guidance, suggestions, and help with minor tasks within the game. The co-pilot AI is often integrated into the gameplay mechanics (not impacting outside of the predetermined rules) to enhance the player's experience, offering tips, strategic advice, or taking over certain routine tasks to allow the player to focus on more complex aspects of the game.
• Interaction: The interaction with a co-pilot AI is usually one-directional, where the AI supports the player's actions and decisions. The co-pilot AI's main goal is to augment the player's abilities and make the game more accessible or enjoyable.
• Examples: AI-driven navigation aids, hint systems in puzzle games, or tactical advisors in strategy games.
  • “Hearing” Professor Oak’s voice in Pokémon instead of an error message
  • Wheatly in Portal 2 (who ironically, is canonically a fictional AI companion)

Co-worker AI Agent

• Role and Function: Co-worker AI agents are designed to work alongside the player or other AI agents, often as equals or peers within the game's context. These agents can perform tasks, participate in missions, and contribute to achieving common goals. Unlike co-pilots, co-worker AIs might not be explicitly designed to guide the player but to collaborate as part of a team or group effort within the game. They also have the ability to be even more reactive, and evolve based on how players interact with them (good and bad).
• Interaction: The interaction with co-worker AI is more reciprocal or collaborative, it can be seen as bi-directional. These agents can take initiative, make decisions, and contribute to the gameplay in ways that are independent of the player's objectives. They can act autonomously to a certain degree and require more sophisticated programming to ensure effective teamwork and communication with the player.
• Examples: Squad members in tactical shooters, partners in cooperative multiplayer games, or any AI character designed to work with the player toward a common objective.
  • Ellie in The Last of Us
  • Atreus in God of War
  • Student Companions in Hogwarts Legacy

The differences between co-pilots and co-workers become apparent in various types of game genres. Each genre has a meta-game that can be benefited or hindered by different types of agents. This makes it important for developers to integrate agents in a way that is cohesive to that experience and does not take away from the core loops that players engage in.

Co-pilots in Game Genres

• Adventure and Role-Playing Games (RPGs): In these genres, co-pilot AIs might manifest as companion characters who provide lore-rich dialogue, hints for quests, or assistance in inventory management. They enhance the narrative experience and help players navigate through complex storylines or puzzles without taking the lead in actions.
• Simulation and Strategy Games: Co-pilot AIs can offer recommendations on resource allocation, suggest strategic moves, or automate certain tasks to allow the player to focus on higher-level planning and decision-making. They act as advisors, enhancing the player's ability to manage complex simulations or strategic battles.
• Racing and Flying Simulators: In these genres, co-pilot AI can assist by handling secondary controls, offering navigational advice, or providing real-time feedback on performance. Their support helps players to focus on the primary aspects of racing or flying, making the experience more immersive and manageable.

Co-workers in Game Genres

• First-person shooters (FPS) and Multiplayer Online Battle Arenas (MOBA): In these competitive genres, co-worker AIs can take on roles equivalent to human players, participating in combat, executing strategies, and adapting to the flow of the game. They are designed to work alongside players (or against them), contributing to team objectives and responding to evolving situations.
• Cooperative Multiplayer Games: Whether a survival game, a puzzle-based adventure, or a cooperative campaign, co-worker AIs fill roles within a team, performing tasks, solving problems, and engaging with the game environment in ways that support the collective goals of the group. They can initiate actions, make decisions, and react to changes in the game state, providing a dynamic and engaging cooperative experience.
• Sports and Team-Based Simulation Games: Co-worker AIs in these genres function as team members, playing according to the rules and strategies of the sport. They make autonomous decisions, execute plays, and adapt to the player's and opponents' actions, contributing to a realistic and competitive gameplay experience.

Takeaway: Co-pilots and co-workers are important distinctions when integrating AI agents into games. Developers will need to be intentional in how they choose to leverage agents based on the specific game to make sure that these agents are complementary to the core game loop and meta-game. In short, co-pilots work with the gamer (like a companion) to problem solve and navigate what is next in the game while co-workers work alongside the player and other AI agents to achieve any number of potential goals (the players, their own, or the enemy’s). These technologies should not be ignored, even with its complexity, as this opens up new possibilities in game design and expands the breadth of gameplay while also allowing players to have more control over the game.