11 Apr 2025

EP4: Bringing History to Life with AI: Inside 1956 Individuals with SKC Games

In this episode, Tim is joined by Asa Burrows and Babita Devi from SKC Games, the team behind 1956 Individuals — AI-powered historical avatars that let people have real conversations with figures like Alan Turing and Queen Eleanor of Castile. Originally a games studio, SKC Games pivoted after discovering how AI could give non-playable characters a life beyond gameplay, evolving into a company focused on AI-driven customer experience for museums, heritage sites, and beyond.

What happens when a games studio starts asking whether a non-playable character could have a life outside the game? For SKC Games, the answer became 1956 Individuals — a range of AI avatars designed to bring historical figures to life through natural, two-way conversation. In a recent episode of the UK AI podcast, Tim spoke with co-founders Asa Burrows and Babita Devi about building historically accurate AI personas, their distinctive approach to AI ethics, and where AI-powered personalisation is headed next.

From Game Design to AI Avatars

SKC Games began during the pandemic as a games company, using AI technologies in character design. That early experimentation led the team to ask a bigger question: could a non-playable character exist beyond the game itself? The result, 1956 Individuals, uses AI avatars to help people access information — and history — in an entirely new way, with a strong focus on customer experience.

Building Historically Accurate, Emotionally Resonant Characters

A standout example of SKC Games' work is its collaboration with Bletchley Park to create an AI avatar of Alan Turing. Rather than relying purely on AI to generate personality and content, the team conducted deep historical research, working directly with Turing's nephew and drawing on archives at King's College. A second example — an avatar of Queen Eleanor of Castile, developed with Leeds Castle — required a different approach, since far less historical documentation exists, meaning historians' interpretations play a larger role in shaping the character's persona.

Facial expressions, mannerisms, and tone of voice are built using a deliberately traditional, human-led process — SKC Games' art team studies photographs, film, and (where possible) direct interaction with living subjects, rather than relying on AI to generate these nuances. AI is instead used to deliver the character: powering natural language processing, custom voice generation, computer vision for object and environment detection, person recognition, and even age detection, allowing avatars to adapt sensitive content depending on whether children are present in an audience.

A Distinctive Approach to AI Ethics

Ethics is a central theme of SKC Games' business, and Asa outlines an addition the company has made to standard AI ethics frameworks: a seventh pillar called "economic prosperity." Alongside common principles like transparency, fairness, and security, this pillar focuses on ensuring that people whose data, voice, or knowledge contributes to an AI character are fairly and continually rewarded — similar to a royalty or Spotify-style model.

The team also emphasises a human, stakeholder-led approach to building historical characters, ensuring the people and institutions connected to a historical figure are consulted and consent to how that person is represented as an AI avatar.

Separating Fact from Opinion — and Avoiding LLM-Generated Answers

One of the most notable aspects of SKC Games' approach is a conscious decision not to let large language models generate factual, historical content. Instead, all backend domain knowledge is vetted by historians, family members, and institutional experts, with LLMs used only to present that verified information in a natural, humanlike way.

This distinction matters most with figures like Alan Turing, where a great deal of public opinion exists alongside verified fact. SKC Games documents its full research and decision-making process — working with curators such as the head curator at Leeds Castle — so that the reasoning behind each piece of content is transparent and traceable, addressing one of the most common criticisms of "black box" AI models.

Copyright, Creativity, and Rewarding Human Contributors

Asa raises a broader concern about the future of AI and creative industries: without fair compensation models, human creators may become less willing to share their work openly, ultimately limiting the richness and diversity of data available to train future AI systems. SKC Games' economic prosperity model is designed to directly counter this risk, ensuring contributors are commercially rewarded whenever their voice, likeness, or knowledge is used — a model the team hopes can scale as 1956 Individuals expands to new historical figures and locations.

The Future: Multilingual, Hyper-Personalised AI Experiences

Looking ahead, Asa and Babita describe a growing range of use cases for AI-powered personalisation, including:

  • Museums and heritage sites — avatars that remember a visitor's interests throughout their visit, offering more relevant information as they move between exhibits
  • Retail and shopping malls — avatars helping visitors navigate quickly and even arrange services like transport
  • Hospitality — conversational check-in experiences that continue beyond the front desk, offering personalised local recommendations
  • Group travel — AI agents capable of balancing the differing interests of multiple people travelling together
  • Multilingual analysis — the ability to gather and analyse visitor insights regardless of the language spoken, helping organisations improve experiences for international audiences.