Here’s what’s on our plate today:

• 🎮 Generative NPCs Rewrite Game Worlds—and Studio Budgets.

• 💸 Figma’s College Dropout CEO Hits Billion-Dollar Design Payday.

• 🤖 OpenAI Tests ‘Do Anything’ Agents for Everyday Tasks.

• 🩺 Chatbot Therapy Gains Traction, Psychologists Weigh the Risks.

Let’s dive in. No floaties needed…

The Laboratory

From smart NPCs to studio shakeups, how AI is redefining the gaming industry

When Rockstar Games released the first trailer of their upcoming GTA 6 game in December 2023, hardly any online news portal skipped the news. News of the delays in its release received similar coverage, pointing to how the industry has shifted from being a niche to becoming a billion-dollar business. Even before the terms ‘artificial intelligence’ or ‘generative AI’ entered public consciousness, the gaming industry was utilizing AI as a key component in video games, including the legendary 1980s Pac-Man, which employed AI to give the ghosts chasing the player distinct personalities.

Cut to 2025, and the billion-dollar industry’s projected market volume of US$733.22 billion by 2030 has made it a prime candidate for implementing more sophisticated generative AI systems that will not only impact how games are developed but also what future games will feel and look like for players.

Rule-based systems metamorphose into lifelike NPCs

In the early days, the use of AI was limited to powering non-player characters (NPCs) using pre-programmed rules and basic pathfinding programs. However, with the advent of more capable systems, game developers began using machine learning, procedural generation, and now generative AI to create more immersive and reactive game worlds.

By 2023, an Australian startup, Replica Studios, had managed to create an AI-powered Smart NPC plugin that could be used by game developers. In its demo, NPC characters could speak using Replica’s AI voices and respond dynamically to player inputs. Behind the scenes, Replica Studios trained NPCs using over 500,000 dialogue lines from 100+ voice actors and integrated OpenAI LLMs (or a user’s model) to generate NPC dialogue in real time, including emotional tone, facial expressions, and responses contextual to gameplay.

The demo offered just a glimpse of how generative AI is being used in game development. According to a report from PC Gamer published in July 2025, one in every five, or around 7% of all the games on Steam, a digital distribution platform for video games, disclosed the use of generative AI in their creation.

Of these, around 60% of the disclosures related to visual assets of all sorts, while other popular uses included audio generation, narrative generation, code generation, and marketing materials.

Big tech’s big dreams

The shift from rule-based systems to games having NPCs with a life-like understanding of the worlds they inhabit has not gone unnoticed. As the technology improves, big tech companies like Google, Microsoft, and Amazon are looking to increase the use of their AI systems to speed adoption.

And it appears to be working. By 2023, Activision Blizzard, the maker of popular titles like Call of Duty, had started using an internal tool called Blizzard Diffusion, inspired by Stability AI’s popular Stable Diffusion image generator.

The New York Times quoted Andrew Guerrero, Blizzard’s vice president of global insights, as saying that “The goal is to remove a repetitive and manual process and enable artists to spend more time on creativity,” Guerrero said in a statement. “Our goal with A.I. has been, and will continue to be, to try to make creative work easier.”

Ubisoft, another popular game developer, also built a tool called Ghostwriter that could produce basic dialogue for games like Assassin’s Creed.

Google is reported to have been working on a new program that it says might eventually replace human play testers with autonomous agents that can run through early builds of a game and discover glitches. Meanwhile, Microsoft has demonstrated adaptive gameplay that uses AI to generate level designs and animations that could take human developers hundreds of hours to produce.

And game studios have taken note. Layoffs have rolled across the industry, with an estimated 6,500 jobs being cut by studios including Amazon Games, Ubisoft, Epic Games, and Niantic.

AI’s impact on jobs in game development

According to estimates, nearly 10,500 people in the gaming industry lost their jobs in 2023, and by 2024, studios had cut more than 11,000 jobs and counting.

A survey from the organizers of the Game Developers Conference found that 49% of the survey’s more than 3,000 respondents said their workplace used AI, and four out of five said they had ethical concerns about its use.

The impact of cuts has been felt in most job titles, like concept artists, graphic designers, asset artists, and illustrators. Yet, as gaming studios turn to generative AI to accelerate development and cut costs, they must also navigate complex copyright laws.

AI raises more questions than it answers

In the U.S., the law states that work seeking copyright must have a human author, and since current AI systems use unlicensed intellectual property, there is an ongoing debate around whether it constitutes copyright infringement or not. This makes the use of generative AI tools for things like concept art, character descriptions, or code problematic, as there are no clear lines around what could constitute infringement.

The use of generative AI in gaming studios and subsequent layoffs have also raised interest in unions. According to a report from Wired, fifty-seven percent of developers surveyed by GDC organizers this year said they were in support of unionizing.

Developers are looking at how Hollywood guild writers moved to secure protections from AI as a roadmap for similar negotiations with their managers about when and how AI could be used. Developers fear that without a proper platform for them to raise their concerns, companies might favor short-term hires to help train AI models, leaving full-time developers in uncertainty.

However, it is not just developers who are wary of generative AI. Gaming studios using AI models beyond the development stage face the additional cost of running these models to power adaptive gameplay and life-like NPCs. The more players engage with AI-powered NPCs, the greater the computing power and cost required to sustain them, and this is considering NPCs do not interact with each other; if that happens, the sheer cost of running them could stretch budgets to a breaking point.

AI innovation vs. industry stability

While players search for realistic worlds inhabited by more interactive NPCs, gaming studios rush to meet their high expectations. The process has long defined the gaming industry; however, now with generative AI, gaming studios have the opportunity to reshape how players interact with the worlds they create. With AI-powered NPCs and adaptive gaming, future games can potentially offer new worlds for gamers. However, studios will have to find the balance between generative AI-powered gaming experiences and business models.

At the same time, studios also have to face questions around copyright infringement, and finding the sweet spot where developers and generative AI can work in tandem to make games more enriching and interactive.

Roko Pro Tip

Prompt Of The Day

Bite-Sized Brains

• College Dropout To Design Billionaire: Figma chief Dylan Field turns a hobby project into a multi-B empire—and hints the next IPO wave will be AI-native design tools.

• OpenAI Unveils ‘Operator’ Do-Anything Agents: A new premium tier trains ChatGPT to book trips, draft code, and run tasks end-to-end so you don’t lift a finger.

• Australia Prescribes Chatbots Over Shrinks: Cash-strapped clinics lean on AI listeners to triage mental-health crises as human counselors hit capacity.

Tuesday Poll

Rate This Edition