Highlights
- EA promotes generative AI to boost efficiency and "rescue" developers from tedious coding work.
- Staff reports significant friction, citing ReefGPT’s broken code and concerns over AI-scripted HR interactions.
- The aggressive automation mandate faces scrutiny ahead of EA’s potential $55 billion acquisition.
EA is pushing artificial intelligence into every corner of its game development pipeline, promising to boost worker efficiency by 30% and rescue burnt-out creators from digital drudgery. But down in the cubicles where the games are actually built, the reality looks far less heroic. Rank-and-file developers are spending their days cleaning up flawed machine-generated code, watching managers use chatbots to script delicate conversations, and quietly wondering if they are being ordered to train their own replacements.
The public champion of this shift is Laura Miele, EA’s President of Enterprise Development. Speaking to industry peers earlier this month at Summer Game Fest, as reported by Eurogamer. Miele insisted the AI rollout isn’t a cold cost-cutting measure but an act of workplace empathy. Her stated mission is to be a "hero" to studio teams by stripping the pure tedium out of their daily workflows.
By letting algorithms handle the mind-numbing grunt work of coding and asset generation, Miele argues that human developers will be freed up to prototype faster, agree on creative directions instantly, and spend their hours doing what they actually signed up to do: make art.
To the company's credit, the raw, mathematical velocity of the technology is undeniable, particularly inside its sports division. Historically, an EA artist spent roughly six grueling months hand-crafting a single digital stadium. Using proprietary generative AI trained strictly on EA’s own legally cleared 3D archives, that timeline collapsed to six weeks, with the company ultimately eyeing a turnaround of just six days.
This supercharged pipeline was the secret engine behind EA Sports College Football 25, allowing the studio to generate thousands of hyper-accurate player likenesses and hundreds of bespoke university arenas that a standard annual release cycle simply couldn't survive. Pointing to wins like this, CEO Andrew Wilson has predicted that more than 50% of EA’s development processes will soon be automated so teams can "get to the fun more quickly."
EA Sports FC 25
ReefGPT: Cleaning Up "Hallucinations & Broken Logic"
Yet, talking to the people holding the mice, getting to the fun feels a lot like doing someone else's chores. Business Insider last October reported that EA leadership pushed its nearly 15,000 employees to adopt AI for almost everything, stretching the mandate into bizarre territory.
Studio managers were reportedly instructed to use algorithms to script real-life, sensitive conversations with their staff regarding pay raises and promotions. On the software side, developers were ordered to feed their own hard-won code into an internal chatbot called ReefGPT, only to watch the tool regularly spit back hallucinations and broken logic that human engineers then had to sit down and manually fix.
This internal friction at EA captures a profound ideological tug-of-war happening across the wider video game landscape. While corporate suites at PlayStation, Capcom, and Epic Games view generative AI as an inevitable industrial revolution, with Epic CEO Tim Sweeney recently declaring the tech will be involved in nearly all future production. Seasoned experts, including former Take-Two AI research head Dr. Luke Dicken, have even warned that the reckless push toward generative tools is actively damaging the broader foundation of computer science.
For the human working inside EA, this grand philosophical debate is attached to a very loud, ticking clock. The entire efficiency push is happening against the backdrop of a historic corporate handover. The publisher is expected to finalize its $55 billion USD sale to a consortium led by Saudi Arabia’s Public Investment Fund. When the ink dries, the new owners will open the ledger, look for that promised 30% leap in speed, and hand EA’s workforce their final answer: they will learn whether the algorithm was brought in to be their assistant or their successor.
