Arc Raiders Studio Defends AI Voice Lines in New Video

Embark Studios has released a video on the sound and music of Arc Raiders, and in it staff explain how the studio uses generative AI, software that creates new audio from patterns in existing recordings, to voice some of the game's characters.

The core problem, according to audio director Bence Pajor, is volume. Each new feature, item and weapon added to the game needs voice lines. Recording every version with human actors would be impractical, and he said actors "would just go crazy if they had to read all of that." Pajor also said the training data used to build the AI voices was made with the voice actors' own performances. Eurogamer reported those remarks.

Two of the game's performers, Jo Rideout, who voices Shani, and Diana Gardner, who voices Celeste, appear in the video to share their views. Gardner described AI use as "freeing", because she could accept other work without worrying about cancelling recording sessions. She also said her Celeste contract "seemed so fair", but she questioned whether she could "take ownership for the final product" of lines created with AI.

Embark's chief commercial officer, Stefan Strandberg, said "we are not in the industry of replacing voice actors", and that the studio should celebrate performance. Strandberg and Pajor acknowledged player criticism, including moments when AI performances broke immersion, the feeling of being absorbed in a game's world. They also pointed to work that is replacing some AI lines.

The debate has followed the game since it came out. Arc Raiders was released in October 2025, and it has received many updates since. PC Gamer reported in March 2026 that the studio's chief executive said a real professional actor is better than AI, after Embark re-recorded some dialogue with human performers. Earlier, in December 2025, actor Neil Newbon criticised the game's AI voice work as "dull as hell" and asked why the studio had not re-recorded lines with actors once the game succeeded.

Earlier reporting also described how Embark used text-to-speech audio, software that turns written text into spoken sound, for many of the game's voice lines. For players, the practical question is what they hear when new content arrives. Embark's video makes the case that human performances sit behind the AI voices, while the performers themselves describe the arrangement in very different terms.