The first consumer industry where autonomous agents will play, trade, and transact is gaming. Through this dedicated department, BGA is making that navigable — for the studios, AI companies, and regulators who will shape this new frontier.






Games are the lowest-stakes, highest-complexity environment humans already share with software. Before agents route your payments at scale, they will play alongside you. A matter of sequencing, not speculation.
Studios don't know which AI companies to talk to. AI companies don't know which studios will ship. Regulators watch from a distance. The rooms exist — they just don't overlap.
For eight years, BGA has organized the movement of autonomous value in games. Now, we are providing the definitive frameworks for Governed Autonomy, Human Accountability, Model-Agnostic Infrastructure, and Agent-Inclusive Economic Design.
Every agent entering a game needs a verifiable identity, scoped permissions, and a reputation that travels with it. Without this, a studio can't tell a trusted agent from a bad actor until it's too late.
Agents will hold value, spend it, and negotiate on a player's behalf. The rails for how they transact — and the guardrails on how they're allowed to behave economically — are being written now.
When an autonomous agent acts inside a game economy, someone has to be able to reconstruct why. Accountability isn't a policy statement — it's a log a regulator or studio can actually read.
An agent's identity, wallet, and reputation shouldn't reset every time it crosses into a new title. Interoperability is what separates one real economy from a hundred isolated experiments.
Studios can't wait for finished law, and regulators can't legislate what they've never seen run. A sandbox lets both sides test agent policy against real gameplay before the rules are fixed.
Frameworks are theory until an agent ships inside a live game. We document what actually happened in production — the deployment, the failure modes, the fix — not the pitch-deck version.
A studio integrating a third-party agent needs a way to trust it without auditing the code line by line. Independent certification is what makes that trust transferable across the industry.
Co-authored alongside ADGM, Teranode, and Deloitte, this report moves past the hype to deliver production-ready frameworks for studios and regulators.
Download the PDF →
A builder-focused side event alongside the world's largest games show, presented with Deloitte. Six companies demoed live — from on-chain game economies to on-device agent runtimes — to a curated room of founders, operators, and researchers. The first gathering of the newly formed Agent AI Council.
CROSS — a game-focused Layer-1 for on-chain economies, powering next-gen blockchain gaming with true ownership and interoperability.
Open-source 3D generative AI on Bittensor — a specialized subnet advancing open-source 3D generation models.
A massively scalable, low-latency execution layer letting agentic AI run thousands of autonomous, data-rich actions in parallel.
Persistent, player-owned sci-fi universes — an AI-driven simulation and extraction shooter set in the Parallel Universe.
A Unity-native AI assistant that indexes entire projects and ships code, shaders, and fixes through agentic automation.
Agents and LLMs running on the player's GPU — dynamic NPCs and real-time world actions with no cloud cost or latency.
Gamescom was the first cohort. Future gatherounds open to companies shipping agent deployments on the report's frameworks. Selected by BGA. Live outcomes, on the record.
First output from the regulatory group. Scoped to agent identity and digital-value definitions in EU / UK consultation.
When memory and context change the next encounter.
Read the comic ↗
Keep your characters when the AI model changes.
Read the comic ↗
How spending permissions keep autonomous agents in check.
Read the comic ↗
[Xsolla]
[Deloitte]
[DMCC]
[Qatar Foundation]Direct from our calendar — no form, no back-and-forth. Pick a slot that works and we confirm within the hour.
Book a slot →