PUBG Ally: A Conversational Embodied Agent as an AI Teammate

Hugging Face Daily Papers Papers
embodied-agent conversational-ai gaming ai-teammate language-model

Summary

The paper introduces PUBG Ally, an embodied AI agent that functions as a voice-enabled teammate in PUBG: BATTLEGROUNDS, integrating language model reasoning with real-time game control for interactive play.

We introduce PUBG Ally, an embodied agent for PUBG: BATTLEGROUNDS that can reason, act autonomously, and play alongside players as a voice-enabled teammate. Building such a teammate requires combining two difficult capabilities: it must perceive and respond to a constantly changing game world under strict latency constraints while interacting naturally with players, keeping its speech synchronized with its actions. Ally therefore combines agentic tool use with real-time game control. A language-model agent uses a controlled interface to inspect game information, interpret player speech, maintain context, decide what to say, and issue high-level action choices that steer a faster control layer for movement, combat, and recovery. Because the player's and Ally's speech and actions continually shape each other and the course of the match, training requires data from actual gameplay. We therefore collect data across nearly 39k sessions in which real players play alongside Ally, recording gameplay, player speech, agent decisions, tool use, actions, and player feedback, and use these records for iterative training. To evaluate teammate quality, we use player feedback and preference comparisons to identify gaps between offline evaluations and player preferences, and iteratively refine the evaluation criteria. Deploying Ally in live service further requires low-latency on-device execution and safeguards for player-facing communication, which we address through model compression, context compaction, targeted safety training, runtime guardrails, and memory redaction. During the live service, we surveyed players in 141 countries. Among respondents whose play with Ally was confirmed in game records, positive responses exceeded negative responses by 25.1 percentage points when asked whether they would recommend Ally, with players describing Ally not only as a tool but also as a teammate or companion.
Original Article
View Cached Full Text

Cached at: 09/25/26, 03:44 AM

Paper page - PUBG Ally: A Conversational Embodied Agent as an AI Teammate

Source: https://huggingface.co/papers/2609.29837 Authors:

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

Abstract

WeintroducePUBGAlly,anembodiedagentforPUBG:BATTLEGROUNDSthatcanreason,actautonomously,andplayalongsideplayersasavoice-enabledteammate.Buildingsuchateammaterequirescombiningtwodifficultcapabilities:itmustperceiveandrespondtoaconstantlychanginggameworldunderstrictlatencyconstraintswhileinteractingnaturallywithplayers,keepingitsspeechsynchronizedwithitsactions.Allythereforecombinesagentictoolusewithreal-timegamecontrol.Alanguage-modelagentusesacontrolledinterfacetoinspectgameinformation,interpretplayerspeech,maintaincontext,decidewhattosay,andissuehigh-levelactionchoicesthatsteerafastercontrollayerformovement,combat,andrecovery.Becausetheplayer’sandAlly’sspeechandactionscontinuallyshapeeachotherandthecourseofthematch,trainingrequiresdatafromactualgameplay.Wethereforecollectdataacrossnearly39ksessionsinwhichrealplayersplayalongsideAlly,recordinggameplay,playerspeech,agentdecisions,tooluse,actions,andplayerfeedback,andusetheserecordsforiterativetraining.Toevaluateteammatequality,weuseplayerfeedbackandpreferencecomparisonstoidentifygapsbetweenofflineevaluationsandplayerpreferences,anditerativelyrefinetheevaluationcriteria.DeployingAllyinliveservicefurtherrequireslow-latencyon-deviceexecutionandsafeguardsforplayer-facingcommunication,whichweaddressthroughmodelcompression,contextcompaction,targetedsafetytraining,runtimeguardrails,andmemoryredaction.Duringtheliveservice,wesurveyedplayersin141countries.AmongrespondentswhoseplaywithAllywasconfirmedingamerecords,positiveresponsesexceedednegativeresponsesby25.1percentagepointswhenaskedwhethertheywouldrecommendAlly,withplayersdescribingAllynotonlyasatoolbutalsoasateammateorcompanion.

View arXiv pageView PDFAdd to collection

Get this paper in your agent:

hf papers read 2609\.29837

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.29837 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.29837 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2609.29837 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

LectūraAgents: A Multi-Agent Framework for Adaptive Personalized AI-Assisted Learning and Embodied Teaching

Hugging Face Daily Papers

LectūraAgents is a multi-agent framework for adaptive personalized learning that mimics professor-student interactions and generates embodied teaching actions aligned with learner profiles. It introduces a hierarchical architecture, an adaptive embodied teaching mechanism, and a Teaching Action-Speech Alignment algorithm, showing consistent improvements over existing approaches.

Creating agent and human collaboration with GPT 4o

OpenAI Blog

Altera, founded by former MIT professor Dr. Robert Yang, launches autonomous AI agents powered by GPT-4o that can play Minecraft collaboratively with humans. The company addresses data degradation in long-duration AI autonomy by combining OpenAI's language models with a brain-inspired parallel architecture.