An introduction to Utility AI
This approach makes our agents behave more naturally than those using other AI techniques 🤩. Utility Intelligence is a Utility AI framework that lets you build emergent, context-aware behaviors for your in-game characters, allowing them to respond intelligently to what’s happening in your game. It plays a key role in the success of modern games, especially now that we’re living in the age of AI, when players expect smarter, more lifelike characters.
- In The Sims (2000) an NPCs current “need” for something (e.g. rest, food, social activity) was combined with a score from an object or activity that could satisfy that same need.
- Just throw in more curves to score each independent consideration and blend all scores together.
- In our experience, these growth rates run well ahead of typical utility IT budget growth, a sign of AI’s strategic centrality.
- Instructions shape behavior; mechanisms change what is physically possible.
- It reflects the massive engineering and organizational work required to move from controlled pilot environments to operational integration.
- For example, in a study, the authors applied a Utility System to calculate the utility values of high-level strategic orders in a team-based tactical game, while Monte Carlo Tree Search (MCTS) was employed to execute these orders at the tactical level.
The high failure rate in production scaling is not an indictment of utility AI or a sign of unsuitability. In our experience, these growth rates run well ahead of typical utility IT budget growth, a sign of AI’s strategic centrality. Estimates vary with scope and come from commercial market-research firms rather than audited data, but the trendlines agree. Regulatory landscape, the agentic turn, connected AI ecosystems, and a practical framework for getting started While Godot comes with plenty of tools to help you understand your game’s behavior and performance, you sometimes need to gather game-specific metrics, and t…
Even something as simple as a defining a set percentage chance for something to happen (e.g. 12% chance to perform Action X) was an early step into utility AI. A behavior can then be selected based on which one scores the highest “utility” or by using those scores to seed the probability distribution for a weighted random selection. Most of the models in the levels were from the free Epic Games content, some assets like shields, bows, ivy and shrubs were modelled in Blender. Possible considerations for ranged combat are the distance to an opponent or the number of arrows. An action consists of any number of considerations, which has an influence on the probability that the AI chooses this action. At any point the need task to use is determined by comparing all Scores and applying the easing curve to it.
Randomizing decisions
These weights can be used to multiply final action scores, even going outside of the range 0-1 to keep our priorities straight. In fact, in Zoo Tycoon 2 in particular, dying is just a bucket that gets priority https://ymlp280.net/the-beginners-guide-to-finding-the-starting-point/ over any other buckets so that an animal won’t accidentally “wake up” at some point. But maybe not all types of actions are always equal, or you want to limit the AI’s options to a certain subset of actions in certain contexts.
Each need node can contain other sub graphs from other systems that delegate the complexity of executing the action that fulfills the need. Utility AI pairs very well with other AI systems, such as Behavior Trees or GOAP. If we were to execute this graph, if the hunger level was high it would likely pick the Cook some Rice task. This feature can save you a lot of time when designing behaviors for your characters.
The result is that the character is selecting the “best” behavior for the given situation at the moment based on how those behaviors are defined mathematically. The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes. Git and GitHub was used as version control system, while YouTrack was used as issue tracker. The algorithms and functions to create and process the Utility AI were implemented in C++ and Blueprints. The decision represents an actual function in the game world, e.g. shooting an arrow. The optimal action that has achieved the highest action score is therefore getting selected.
- When it’s time to score our actions, we’ll multiply the raw action scores by each weight, which will result in doubling the score for attacks and make them that much more likely to be selected.
- Through the implementation of the Utility AI system in Unreal Engine, it turned out that a hybrid of this approach and behavior trees can achieve an optimal result.
- To properly score our healing action, let’s first look the utility of healing based on how low on health we are.
- The combinations of these values gave a score to the action that told the Sim what it should do.
- The AI will simply say “this is the action to take” and then your game code is responsible for parsing that into actionable data however you see fit, making it a highly flexible system.
- Want to consider more inputs, such as the action point cost of different actions, cooldown times, etc?
With Unto Deepest Depths now out, let’s talk about the game’s development cycle, how it changed along the way, and lessons learned. We could assign a weight of two to offensive abilities and a weight of one to all others. Say we want to have an aggressive enemy that strongly favors attacks over defensive or idle actions unless the situation is really dire.
This phase identifies specific use cases and measures early ROI. Banking routes customer financial data through AI systems daily. Utility leaders frequently express concern that external AI systems could expose proprietary or sensitive operational data.
We could say that any action within 0.1 of the top scoring action is an acceptable alternative, resulting in three attacks to choose from and adding a bit of variety to the gameplay. Rather than always picking the top scoring action, randomly choosing any action within some percentage of the top scoring action can add variety to the AI while still keeping it on the right track. But how do you decide what score, exactly, to assign to an action when there are multiple things to consider? It’s also worth noting that Utility AI is not concerned with how to take a given action, only what action it should take, meaning that the concepts behind it are fully compatible with whatever technique you’re using to execute actions in your game, whether that be via state machine or some other mechanism. Should a unit run up to an opponent to deal strong melee damage, but risk exposing themselves, or should they hang back in cover and opt a weaker ranged attack? And considering there are typically multiple ways to satisfy each need, such as making a meal versus ordering pizza, which specific action is the “right” choice here?
The Connected Enterprise: MCP and the Internal AI Ecosystem
Year 2-3 expands the MCP ecosystem toward integrated operations, and Tier 4 (self-hosted SLMs) now enters the conversation earlier for utilities with defined sovereignty needs, at appliance-scale cost. (Cost and timeline figures in this section are illustrative planning ranges, not vendor quotes or industry benchmarks; verify current pricing and scope-specific estimates before budgeting.) Published routing research reports cost reductions of roughly 35% to more than 85% depending on the task, while retaining about 95% of the larger model’s quality.
Now, when health is relatively high, this consideration will rank healing actions low, only scoring such actions at 0.25 when at half health, but as health decreases further, our score will quickly ramp up, scoring healing actions at 0.5 when at 25% health, and 0.75 when at 10% health. Instead, we may want a curve that scores healing when at high health lower, but really ramps up things once we get to a critical point. We could start out by defining a simple inverse linear curve where we get a score of 0 when at max health and 1 when on the verge of death, meaning we won’t even try to heal when at max health but will almost certainly decide to do so when we’re on our last few hit points. Zoo Tycoon 2 used it for deciding both animals and guest actions, changing what actions were considered depending on the current context, and The Sims series uses it, along with some extra design considerations I’ll come back to in a moment, to decide what action a Sim should take when left to their own devices. Utility AI can also be integrated with other architectures in a modular approach, where each architecture handles a specific task.
The AI Deployment Spectrum
In our view, banks have been able to scale AI in part because their governance frameworks are mature. A $5,000 appliance running an openly licensed, fine-tuned model, behind the utility’s own firewall and trained on its own documents, puts genuinely private AI within reach of a municipal utility or rural cooperative. For a decade, “data integration” meant warehouse projects that consumed budgets and produced interactive dashboards. Utilities can capture most of the ecosystem’s value (planning, analytics, maintenance prioritization, customer operations) without touching the control plane. For utilities, the connected-enterprise path runs through the IT domain and stops at the OT boundary. Instructions shape https://expandsuccess.org/creating-a-budget-that-works/ behavior; mechanisms change what is physically possible.
