Gemini · Google · AI Safety · Decrypt
Google Says AI Isn't Building Itself Yet—Even Though It Writes 75% of the Code
Compiled by KHAO Editorial — aggregated from 1 source. See llms.txt for citation guidance.
★ Tier-1 Source
Google DeepMind went looking for proof that AI is speeding up the development of AI itself.
Key facts
- 1/9 In the Gemini 4 Argon Safety & Security Report, they published their first holistic empirical assessment of ML R&D automation and acceleration inside Google DeepMind, combining capability benchmarks
- Most production code at DeepMind is now more than 75% AI-written, and the share of fully delegated pull requests (bundles of code changes submitted for review before they go live) rose from 7.5%
- To measure progress, the team built GCI v1, an internal index that blends the scores of 123 AI models across 92 benchmarks into a single number per model—think of it as an IQ score for AI
- In the report’s chart of that index, Gemini 1.5 Pro sits near 0.5 points in late 2024 and Gemini 3.8 Flash above 2.9 in September 2026, with nearly every release hugging a straight line
Summary
Google DeepMind found no statistically significant acceleration in the pace of its AI progress after April 1, 2026; its capability index has climbed a steady 0.1 points a month. Most production code at the lab is now more than 75% AI-written, but about 88% of code changes still need humans to steer the AI or write code themselves. Gemini 4 Argon solved roughly half of the internal research engineering tasks it finished, far below the 90% score DeepMind treats as a red flag for full automation. In the safety report for Gemini 4 Argon, its most powerful model, the lab concluded AI research isn’t moving any faster, even though AI now writes most of the code DeepMind’s engineers ship.