AI Policy Pulse
←Home

AI openness: Balancing innovation, transparency and risk in open-weight models

OECD AI Policy Observatory OECD Published: 2025-08-28 10:14:43

OpenAI introduced GPT-OSS, a series of open-weight AI models, in August 2025 to increase transparency and innovation in the AI community, sparking debates about the openness of such models. This move prompted discussions on finding a balance between innovation, transparency, and potential risks associated with sharing AI models.

Full text
<p>In August 2025, OpenAI announced GPT-OSS, a family of open-weight models that provide public access to the trained parameters of a frontier-level AI system. This gives a new sense of urgency to the debate over how “open” artificial intelligence models should be. Some heralded the move as a victory for transparency and innovation. Others criticised [&#8230;]</p> <p>The post <a href="https://wp.oecd.ai/balancing-innovation-transparency-and-risk-in-open-weight-models/">AI openness: Balancing innovation, transparency and risk in open-weight models</a> appeared first on <a href="https://wp.oecd.ai">OECD.AI</a>.</p>
Recommended reads
A five-step roadmap for closing the AI evaluation gap to strengthen trust, security, adoption and effective AI governance. The post A five-step roadmap to closing the AI evaluation gap appeared first …
OECD AI Policy Observatory · OECD 2026-07-31
How Salesforce sees the HAIP Reporting Framework reducing AI governance fragmentation and making transparency a competitive advantage. The post HAIP is transforming transparency from a compliance burd…
OECD AI Policy Observatory · OECD 2026-07-16
How the UK’s ATRS strengthens algorithmic transparency, public trust and accountability in government AI. The post Designing transparency for government AI: Insights from the UK’s Algorithmic Transpar…
OECD AI Policy Observatory · OECD 2026-04-14