OpenAI to stop training AI models on the open internet
OpenAI is ending the practice of training its most advanced AI models directly on the open internet. The company cites significant risks as the reason for the decision.
What happened?
OpenAI has announced that it will stop training its most advanced AI models directly against the open internet. The company states that the method entails "significant risks". This decision means that future frontier models will be disconnected from real-time data and open web content during the training phase.
Key facts
| Beslut | Stopp för modellträning på levande internet |
|---|---|
| Orsak | Betydande risker ("significant risks") |
| Datum för besked | 23 oktober 2026 |
Why it matters
The decision breaks with previous practices where large-scale scraping of live web data was central to the development of frontier models. The fact that a leading AI company is voluntarily opting out of direct internet training highlights growing internal security and quality trade-offs surrounding the training process.
Who is affected?
The change affects AI developers and companies building solutions on top of OpenAI's frontier models that are dependent on how training data is collected. It also impacts users of future models through potential changes in the models' training methodologies.
Impact on the EU
The decision is global and covers all markets where OpenAI's models are trained and deployed, including the European Union. However, the change does not immediately alter the availability of existing OpenAI services within the EU.
What else you should know
The announcement marks a strategic shift, coming just weeks after an OpenAI engineer publicly argued that such a disconnection would not be a viable solution. The company has not specified exactly which risks are involved beyond the description that they are considered significant.
Quick answers about this story
Vad har hänt?
När hände det?
Varför spelar det roll?
The link opens in a new window and leads to the publisher's own site.
Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.
AI-verktyg i artikeln
Topics
Get similar news straight to your inbox
The reader's room
Send in a question or an addition. The newsroom reads everything before it's published and replies when relevant. No AI-generated text – just people.
Sign in to submit a comment or question.
Read the article through your role
- Decide whether this affects strategy over 6–12 months or is just noise.
- Discuss with leadership: do we own the right question or does ownership need to move?
- Ask: what risk are we taking by NOT acting on this this quarter?
Generated angle — not editorial analysis of "OpenAI to stop training AI models on the open internet"