Critical Criteria

The Synactors Guild Awards exist to recognise excellence in synthetic performance, and that intention is only as meaningful as the criteria used to judge it. This page sets out how the guild's reviewers approach the work of evaluation: what we look for, what we deliberately do not look for, and how we handle the questions of authorship and creative agency that have become increasingly pressing as the technology develops.

These criteria are published not as a settled code but as a working document. They will be revised as our understanding of the form develops, as new kinds of synthetic performance emerge, and as the critical vocabulary the guild is in the process of building becomes more precise.

Performance, not production

The first and most important criterion is also the most basic: the guild evaluates performance, not production. We are assessing what a synthetic character does in the moment of its encounter with an audience — the quality and coherence of its expressive presence — not the sophistication of the tools used to create it, the scale of the studio that made it, or the commercial success of the production in which it appears. A technically modest character that produces a genuine and surprising emotional effect will always interest us more than a photorealistic figure that produces none.

This distinction matters because the existing awards landscape persistently conflates the two. Technical achievement awards, visual effects categories, and development citations all have their place; they simply are not what the SyGAs are for. Where the guild departs most sharply from other bodies is in its insistence that what a synactor does constitutes a performance in a meaningful sense, and that this performance can be evaluated with the same critical seriousness we bring to any other kind of performed art.

Fidelity versus expressivity

Visual fidelity — how closely a synthetic character resembles a human being in its surface appearance — is one of the most seductive and most misleading measures available to a critic of synthetic performance. It is seductive because it is immediately visible and immediately impressive. It is misleading because it is entirely possible to produce a digital figure of extraordinary surface realism that cannot sustain a performance of any depth, and entirely possible to produce a stylised or technically imperfect figure whose expressive range is remarkable.

The guild does not ignore visual fidelity, but we do not treat it as primary. What we ask of a synthetic character's appearance is whether it is consistent with and in service of the performance — whether the character looks the way it needs to look in order to do what it does. A deliberate stylisation may be precisely right for a particular performance register. An unresolved inconsistency between visual systems — a character rendered at high fidelity in one context and at a noticeably lower resolution in another — is interesting to us not as a technical failure to be noted and dismissed, but as something to understand: Is it a production compromise? A deliberate choice? A symptom of where the creative investment was concentrated? The answer affects our reading of the work.

What we seek, and weight most heavily, is expressivity: the capacity of the synthetic character to convey interior states through the full range of available means — movement, posture, vocal quality, timing, response to other characters and to the environment. A performance that surprises us, that does something we did not expect, that reveals a dimension of the character we had not anticipated, is the performance we are looking for.

The interactive dimension

Synthetic performance in games poses critical challenges that performance in linear media does not. A character in a film delivers one performance; a character in a game delivers, in effect, a different performance in response to each player, shaped by choices the designers cannot fully predict. The guild's reviewers are therefore required to engage with a game performance across multiple playthroughs and across the full range of the character's expressive states — not merely the moments of greatest dramatic intensity, but the transitions, the pauses, the ambient expressions of character that accumulate into a coherent presence over the duration of play.

We also attend to the relationship between synthetic performance and player agency. A character whose performance is entirely independent of the player's actions may be impressive as an acting achievement but limited as a game performance; a character who responds to and is genuinely shaped by the player's choices, without losing expressive coherence, demonstrates a kind of craft specific to the interactive form. The guild regards this as one of the most demanding and least recognised achievements in synthetic performance.

Creative agency and the question of authorship

The guild's position on questions of creative authorship is deliberately distinct from the position taken by bodies such as the Academy of Motion Picture Arts and Sciences, whose 2026 rules exclude from eligibility any performance not demonstrably carried out by a living human being. We understand the reasons for that exclusion, and we do not dispute its appropriateness for the Academy's purposes. Our purposes are different.

The Synactors Guild was founded on the premise that synthetic performance is a form of performance, and that the question of whether a human being carried out a specific physical act is less important, for our purposes, than the question of whether a performance of genuine quality was achieved. We are in the business of evaluating what was made, not policing how it was made. Consequently, we do not disqualify a nominated performance on the grounds that it was substantially shaped by an AI system, generated algorithmically, or produced without a human performer as its primary foundation.

What we do require — and this is the critical discipline our position demands — is that the reviewer describes, with precision, where the creative agency in the performance appears to lie. This is not a simple task. The landscape of creative contribution in synthetic performance is rarely clear-cut: a performance may have been designed by a human director, generated by an AI system, refined by a human animator, and shaped at the moment of play by a player whose choices were anticipated only probabilistically. The reviewer's job is not to award a single credit but to account honestly for this complexity. An award citation from the guild should tell the reader something true about how the performance came to exist, not simply celebrate its existence.

We hold this position not because we are indifferent to questions of human creative labour — we are not — but because we believe the alternative leads to an increasingly untenable critical position as the technology develops. A framework that defines synthetic performance as always and necessarily human-authored cannot survive contact with a field in which the boundaries of human authorship are actively and rapidly shifting. The guild chooses to engage with that shift directly rather than to defer it.

The world as performer

The guild's categories include recognition for synthetic character performance — the individual or ensemble synactor. They also include, as a distinct category, recognition for the game or animated world itself as a performing entity. This requires some explanation, since it departs from the conventions of most critical frameworks.

A game world is not merely a backdrop. Its architecture, its weather, the behaviour of its crowds and creatures, the way it registers damage or growth or the passage of time — these are expressive acts. A world that communicates the emotional register of its narrative through spatial design, through the feel of movement within it, through the way it responds to what the player does, is performing in precisely the sense the guild means. It is conveying something that could not be conveyed by a character alone. The destroyed cities of a post-apocalyptic future, the oppressive geometry of a dictator's capital, the particular quality of light in a world that is dying — these are performances that deserve critical attention on their own terms, not merely as context for the characters within them.

In procedurally generated and dynamically responsive worlds this becomes even clearer. A world that generates itself in response to the player's presence, that produces encounters and environments that were not individually authored, is a synthetic performer in the most literal sense: it is creating expressive content in real time, from a system rather than a script. The guild regards the evaluation of such worlds as an extension of its core critical project, not a departure from it.

The criterion for this category is the same as for character performance: we ask whether the world does something expressive that could not have been achieved otherwise, whether its behaviour is consistent and in service of a legible aesthetic intention, and whether it produces responses in the audience that go beyond the merely functional. A world that makes the player feel something the characters alone could not produce is a world that has performed.

The adaptive and procedural score

Music that responds — that changes in real time according to the player's state, that generates rather than plays back, that creates emotional atmosphere for a moment no composer specifically prescribed — is synthetic performance in perhaps the purest sense the guild recognises. The composer authors a system; the system performs. The distinction from a fixed composed score is not one of quality but of kind: an adaptive score is doing something a fixed score cannot, and that something is exactly what the guild exists to evaluate.

The history of adaptive game music is longer than is generally recognised. Dynamic layering systems, branching musical structures, and procedurally generated accompaniment have been present in games since the early 1990s. What they share is the displacement of the single authored performance by a generative one: the music you hear is not the music anyone composed for that moment, because no one knew that moment would occur in precisely that form. The guild regards this as a creative achievement of the first order, and one that has been almost entirely invisible to existing awards frameworks, which evaluate composed scores and treat adaptive systems as technical rather than artistic achievements.

The criterion for this category is whether the musical system produces expressive results that are genuinely responsive to their context — whether the music performs, in the sense of doing something in relation to what is happening that a fixed cue could not do. We are not evaluating the quality of the composed elements in isolation but the quality of the system's expressive behaviour as a whole. A score that surprises us, that produces a musical moment we could not have anticipated and that is precisely right for the moment it accompanies, is the score we are looking for.

What we are waiting for

The most honest statement we can make about our critical criteria is that they are oriented towards a kind of performance we have not yet seen in its fullest form. The guild was founded in anticipation of synthetic performers capable of something genuinely autonomous — not the removal of the human from the process, which is neither possible nor desirable, since the human is always at minimum the audience, and the synthetic performer is itself a human construct existing within and for the human world — but the emergence of performances in which the creative intelligence is substantially the synactor's own.

We are not certain how close that prospect is. What we observe in the best current synthetic performances is a complex and often asymmetric distribution of creative contribution: a human team whose investment is visible in certain dimensions of the work, and an AI system or generative process whose contribution is visible in others. The uneven texture of contemporary synthetic performance — the extraordinary in one register, the unresolved in another — is itself evidence of a practice in transition. The guild reads those unevennesses not as failures to be discounted but as data about where we are and where the art form is going. They are, in their way, encouraging.

The criteria above will serve us for now. We expect to revise them.