AI Voices, Human Performances: The Arc Raiders Debate

The release of Arc Raiders, a sci-fi extraction shooter from Embark Studios, triggered an intense debate about the role of AI-generated voices in video games. What started as a design and production choice has rapidly become a lightning rod for conversations about ethics, creativity, labor, and the future of performance in interactive media.

This article explores what happened with Arc Raiders, why AI voices are so contentious, and how developers, actors, and players are thinking about the technology’s role in the next generation of games.

How Arc Raiders Landed in the Middle of the AI Voice Firestorm

Arc Raiders launched into a market already primed for controversy around generative AI. The game features a mix of traditional recorded performances and text-to-speech (TTS) AI voices for various non-player characters (NPCs). While the core concept and gameplay attracted interest, critics and players quickly honed in on the AI voice work.

Several factors fueled the backlash:

• Media reviews that specifically called out AI-generated lines as immersion-breaking.
• Player feedback highlighting unnatural cadence, odd emphasis, or flat delivery in some NPC dialogue.
• Ongoing industry-wide concerns about AI tools replacing or undercutting creative workers, especially actors and writers.

Embark Studios has previously acknowledged using AI-assisted voices in its earlier title The Finals. In public statements, the studio framed AI as a tool for flexibility and iteration, not as a replacement for actors. For example, in a statement shared with major games press, Embark explained that it uses a combination of recorded voice audio and TTS depending on context, and that making games without actors is not an “end goal.”¹

Despite this, Arc Raiders became a focal point because it arrived at a moment when both workers’ unions and players are far more aware of how generative AI is being deployed in entertainment.

AI as a “Production Tool,” Not a “Creative Voice”

One of the sharpest reactions to the Arc Raiders situation has come from other developers. Creators behind the upcoming narrative game Dispatch have argued that AI voices are, at best, a production solution—useful for speed and logistics, but fundamentally limited as an instrument of creativity.

Their position reflects a broader sentiment among narrative designers and directors:

• AI voices can quickly produce placeholder dialogue or scratch audio.
• They make it cheap to try new lines, scenes, or variations before final recording.
• But they struggle to provide genuinely novel, emotionally specific performances.

From this perspective, TTS is closer to a sophisticated spreadsheet than a creative collaborator. It can recombine patterns it has seen in training data but cannot truly originate a performance in the way a skilled actor can.

Why Human Performances Still Matter

Game developers who favor human actors often point to several advantages of live performance:

• Spontaneity and chemistry: When actors record together, they influence each other’s timing, tone, and emotional choices in ways that are difficult for AI to mimic.
• Interpretation: Actors interpret subtext, tone shifts, and character history, often adding nuance that is not explicitly in the script.
• Iteration with writers and directors: In the studio, scripts can evolve as actors and writers discover better lines, reactions, or characterization.

Ultimately, supporters of human-led performance argue that these qualities are what make characters memorable and believable—particularly in story-driven games or titles that rely on strong atmospheric immersion.

Why Studios Are Turning to AI Voices Anyway

If AI voices are so controversial, why use them at all? For many studios, the answer lies in the production pipeline. Modern games, especially open-world or systemic titles, can contain tens of thousands of voiced lines: incidental barks, tutorial prompts, dynamic responses, and localized variations.

For a studio under time and budget pressure, AI can be attractive for:

• Filling in low-priority or low-emotion lines.
• Rapidly iterating placeholder VO while scripting is in flux.
• Supporting live-service games that need frequent content updates.

Embark has described its use of TTS in similar terms—particularly for contextual in-game callouts and other context-heavy lines that might be too expensive to record at scale with traditional methods.¹

The Ethical Fault Lines: Consent, Compensation, and Control

Beyond questions of quality and creativity, AI voices raise a dense cluster of ethical and legal issues. These concerns are not theoretical: major entertainment unions have already addressed them directly.

Performers’ Unions Push Back

The Screen Actors Guild – American Federation of Television and Radio Artists (SAG-AFTRA), which represents many voice actors, has been explicit about the risks of AI in performance. The union’s guidance warns of scenarios where an actor’s performance is captured, modeled, and reused—or even altered—without adequate control or compensation, particularly as AI tools become more powerful.²

As a result, union contracts increasingly seek:

• Clear consent for any AI training or cloning of a performer’s voice or likeness.
• Limits on reuse of recorded material beyond its original context.
• Additional compensation for derivative works generated from an actor’s performance.

Law and Policy Are Catching Up

Regulators are also starting to respond. For example, the U.S. Federal Trade Commission (FTC) has issued guidance warning companies that generative AI can be used deceptively and that businesses remain responsible when AI tools facilitate fraud or misuse.³ Likewise, the European Union’s AI Act includes rules requiring transparency when people interact with AI systems, and places additional obligations on certain uses of generative AI.⁴

While these policies are aimed broadly at AI, they set a precedent that will shape how voice technologies are deployed in media and games. Studios relying on AI voice models trained on human performance will face increasing scrutiny over consent, data provenance, and disclosure.

Player Immersion vs. Technical Efficiency

At the heart of the Arc Raiders debate is a simple experiential question: do AI voices feel right in a high-budget game? Many players argue that, even when technically impressive, AI dialogue often has subtle giveaways:

• Strange emphasis on certain words.
• Slightly off timing between pauses.
• Limited ability to convey complex emotions like grief, fear, or layered sarcasm.

These small issues can be enough to pull a player out of the world, especially in cinematic or narrative-heavy experiences. For some, this is a deal-breaker; for others, it may be acceptable in non-critical roles like generic enemies or automated announcements.

Critics of Arc Raiders’ implementation have argued that AI voices were used in ways that were too visible, too frequently, or too close to human characters that players are expected to care about. Combined with broader industry concerns about AI, this magnified what might otherwise have been viewed as a technical compromise.

Industry Perspectives: Where Developers Draw the Line

Developers and studios are far from unanimous in their attitudes to AI voices, but several common positions are emerging:

1. AI for Prototyping Only

Some teams embrace AI voices as temporary tools during development and then replace everything with human actors before launch. In this model:

• Designers can hear their dialogue in context early.
• Iteration is cheaper and faster.
• Final performances remain fully human.

This approach tends to be viewed as more ethically acceptable, assuming no unauthorized cloning is involved and that actors record only the final approved scripts.

2. Hybrid Production: Human Leads, AI Background

Another common stance is a hybrid approach, where primary characters and story-critical scenes are performed by actors, while AI covers incidental dialogue, radio chatter, or background NPC lines. Proponents argue this can preserve emotional depth while controlling costs for large-scale titles.

However, the Arc Raiders backlash suggests that players are becoming more sensitive to this trade-off, especially when AI voices are easy to spot or feel out of place.

3. Full Embrace of AI Voices

A smaller but vocal group of technologists see AI voices as the future default, arguing that as models improve, they will reach parity with human performance for many use cases. They point to rapid advances in speech synthesis and emotional control, often backed by research from large technology firms and academic labs.⁵

Still, even some technically optimistic developers acknowledge that reaching true emotional equivalence—and addressing ethical concerns—will require more than just better neural networks.

What Players Are Asking For

Player reactions to Arc Raiders and similar cases reveal several recurring expectations:

• Transparency: Many players want studios to be upfront about when and where AI voices are used.
• Quality thresholds: If AI dialogue cannot match a minimum quality bar, they would prefer fewer lines or a different design approach.
• Respect for workers: A growing segment of the audience cares about whether games are made under fair conditions, including how actors and other creatives are treated and compensated.
• Player choice in some contexts: Some have suggested options such as toggling AI-heavy commentary packs on or off in multiplayer games.

In the case of Arc Raiders, mounting criticism from media and community discussions helped push AI voices in games into the mainstream spotlight, strengthening calls for clearer policies and higher standards.

Practical Guidelines for Responsible AI Voice Use

The Arc Raiders controversy is likely not the last such flashpoint. Studios considering AI voices can reduce both ethical risk and player backlash by adopting some practical principles:

• Obtain explicit consent: If a model is based on a specific actor, ensure contracts clearly cover AI training, usage, and rights.
• Prioritize human-led performances for key roles: Story-critical characters should generally be fully human-performed.
• Be transparent: Include credits that disclose when and how AI tools were used, and communicate openly in marketing and patch notes.
• Involve actors in the process: Some performers may be open to licensing digital versions of their voices under fair terms; their input can shape better practices.
• Monitor player feedback: Treat AI voice usage like any other design choice—subject to iteration, rollback, or replacement if it harms the experience.

Handled carefully, AI voices can remain what some directors call them: a production tool rather than a creative replacement. Mishandled, they risk undermining trust with both audiences and the people whose voices make game worlds feel alive.

FAQs: AI Voices in Games and the Arc Raiders Controversy

What exactly did Arc Raiders do with AI voices?

Arc Raiders uses a mixture of human-recorded performances and AI-generated text-to-speech for certain NPCs and incidental dialogue. While high-profile scenes rely on actors, some in-game lines are delivered by AI voice models, contributing to criticism that the overall experience feels uneven.

Are AI voices replacing human actors entirely?

In most mainstream game productions, no. AI is currently used primarily for prototyping, background chatter, or large volumes of minor lines. Major roles in story-driven games still rely heavily on human actors, and unions like SAG-AFTRA are actively negotiating to prevent unapproved substitution of AI for human work.²

Why are voice actors concerned about AI?

Voice actors worry that their performances might be recorded, modeled, and reused by AI systems without proper consent or compensation. They also fear downward pressure on wages if studios can cheaply generate synthetic performances. These concerns mirror wider debates in film, TV, and other creative industries.

Is it legal to train AI on an actor’s voice?

The legality depends on jurisdiction, contracts, and how the training data was obtained. Some regions recognize strong rights to one’s voice and likeness, while others are still developing explicit rules. Regulatory bodies like the FTC and frameworks such as the EU AI Act are beginning to define boundaries around deceptive or unauthorized uses of generative AI.^3,4

Will AI voices eventually sound as good as human performances?

Technically, AI voices have made rapid progress and can already produce natural-sounding speech with controllable tone and emotion. Research from major tech companies shows continual improvements in expressiveness and control.⁵ However, reaching the depth, spontaneity, and interpretive nuance of top-tier human performances—and doing so ethically—remains a significant challenge.

Looking Ahead: Lessons from Arc Raiders

Arc Raiders has become a case study in how not just technology, but also context and communication, shape public reaction. The game arrived at a moment when players, critics, and workers are all more attuned to the impact of generative AI on creative industries.

Several key lessons are emerging:

• AI voices will be judged not only on technical quality but on their ethical and creative implications.
• Studios that treat AI as a quiet shortcut risk reputational damage if players perceive the result as hollow or exploitative.
• Clear policies, strong labor protections, and a focus on human-led creativity can make AI tools more acceptable to both audiences and workers.

For now, the safest path forward for many developers may be to embrace AI voices as supporting infrastructure—a way to speed up iteration, fill non-essential lines, and experiment—while keeping the heart of their stories firmly in the hands of human performers.