adminAI-powered NPCs could make game conversations feel more open-ended, giving players room to ask their own questions and shape an investigation. A University of Bristol study with developer Meaning Machine examined how players responded to this kind of interaction in the murder mystery game Dead Meat. Most participants described the experience as enjoyable and rewarding, though some also found its open-ended nature overwhelming.
Image credit: Meaning Machine
What the study found about AI-powered NPCs
In Dead Meat, players take on the role of investigators and question suspects. The study report describes 68 participants and lists 31 male, 31 female, and seven non-binary participants. Researchers reported that 95% found the experience enjoyable, while 97% said playing the game was rewarding.
Three-quarters of participants felt that the game let them express themselves or make meaningful choices. Researchers said these responses were consistent across the different game types included in the study. The results suggest that players may value AI not simply as a way to automate existing game elements, but as a way to support different kinds of interaction.
Practical takeaway: For players, an AI-driven conversation can offer more ways to approach an investigation. For game designers, the findings point to the importance of giving players freedom while still helping them understand what they can do next.
Player freedom can also feel overwhelming
Some participants appreciated being able to shape the interrogation themselves. One described the experience as rewarding because they could make up their own questions. That freedom, however, did not feel comfortable to everyone at every moment.
One player said the freedom could feel exhilarating, immersive, and exciting, but could also become overwhelming. This highlights a design challenge: open-ended dialogue can encourage expression, but players may need cues or a clear sense of purpose to avoid feeling lost.
What to look for: In games with conversational NPCs, notice whether the game gives enough context to help you decide what to ask or do. A useful balance can let players experiment without making the objective difficult to follow.
How the findings fit into the wider AI debate
Lead researcher Dr Richard Cole said the research centers players’ feelings about AI-powered experiences in practice, rather than treating the debate as purely theoretical. Meaning Machine co-founder Thomas Keane also distinguished between AI that replaces creative work and AI used to create new forms of interaction.
That distinction matters because the use of AI in game development has prompted criticism, including concerns about how generated voice lines may affect performers. The study focuses on a different question: how players respond when AI supports conversations and choices during play. Its results offer evidence about the participants’ experience, but should not be treated as proof that every player or game will respond in the same way.
What comes next for AI-powered NPCs
A full research paper was expected at the end of the year and was set to include a similar analysis of players using Meaning Machine’s later game, Blood Will Out. Researchers collected the same volume of data for that title, which uses a more advanced version of the studio’s technology. Comparing the games may help clarify how changes to the system affect player experience.
Overall, this study suggests that AI-powered NPCs can support enjoyable, rewarding play when they enable meaningful choices, while also showing why clear direction remains important. Players interested in these systems should consider both sides: the freedom to interact in their own way and the guidance that helps make that freedom manageable.
