# Anthropic’s Guillaume Princen Says Open-Weight Models Carry More Risk by Definition

Times of AI Desk · 2026-10-14 · Policy

[https://timesof.ai/2026/10/anthropic-princen-open-weights-higher-risk](https://timesof.ai/2026/10/anthropic-princen-open-weights-higher-risk)

> At Ireland’s International AI Summit in Dublin on Oct. 14, Anthropic’s head of international, Guillaume Princen, told Bloomberg TV that open alternatives to Anthropic’s models carry “more risk” by definition. The Next Web reads the line as broader than CEO Dario Amodei’s July position. Bloomberg’s report was read via Business Standard syndication.

Anthropic’s commercial lead outside the United States has drawn what The Next Web calls a broader line on open weights than the company’s July policy post. **Guillaume Princen**, Anthropic’s head of international, told **Bloomberg TV** at the **International AI Summit** in **Dublin** on **Wednesday, October 14**, that open alternatives to Anthropic’s models carry more risk. **Business Standard** carried Bloomberg’s report; **The Next Web** also covered the remarks.

“We need to be really careful with open-source on safety,” Princen said, according to Bloomberg. “I can’t talk for open models. But by definition, there’s more risk with them.”

## How that sits next to Anthropic’s July line

In [July](https://timesof.ai/2026/07/anthropic-position-open-weights-models), CEO **Dario Amodei** wrote that Anthropic has never advocated banning open-weight models and that open-weights models without dangerous capabilities are a public good. The Next Web reads Princen’s “by definition” line as categorical where Amodei’s was conditional on capability. The interview is not a change to Anthropic’s Responsible Scaling Policy or product access rules.

Bloomberg places the comments amid competition from open-weight models and notes that France’s **Mistral** recently released an open-weight model as a sovereign option for Europe. Mistral’s [Large 4 preview](https://timesof.ai/2026/10/mistral-large-4-preview-1t-open-weights), announced October 6, is covered separately. Bloomberg also ties Princen’s concern to Anthropic’s own recent research on small models that could self-replicate across cheap internet-connected devices — covered [here](https://timesof.ai/2026/10/anthropic-frontier-red-team-self-replicating-ai-iot-devices).

## Price claim and what he declined to say

Princen said Opus 5.5 is **40% cheaper** than a prior version, Bloomberg reported. The Next Web says that is a typical-cost figure: the list price fell **20%**, with the rest from using fewer tokens. Princen declined to comment on Anthropic’s IPO plans or training spend, Bloomberg said.

## Limits

- Primary is a Bloomberg TV interview read through Business Standard syndication and The Next Web; not a company blog post.
- “More risk by definition” is Princen’s phrasing in an interview, not a published Anthropic policy update.
- Opus 5.5 cost figures are attributed as above; list prices were not re-checked for this article.

## Sources

- [Business Standard (Bloomberg): Open-source AI models carry more cybersecurity risks: Anthropic executive (October 14, 2026)](https://www.business-standard.com/technology/artificial-intelligence/open-source-ai-models-carry-more-cybersecurity-risks-anthropic-executive-126101400731_1.html)
- [The Next Web: Anthropic’s international chief says open models carry more risk by definition (October 14, 2026)](https://thenextweb.com/news/anthropic-princen-open-weights-risk)
- [Anthropic: Our position on open-weights models (July 27, 2026)](https://www.anthropic.com/news/position-open-weights-models)
- [Anthropic: AI models may be capable of self-replicating without access to a datacenter (October 12, 2026)](https://www.anthropic.com/research/self-replicating-ai)
- [Mistral AI: Introducing Mistral Large 4 (October 6, 2026)](https://mistral.ai/news/mistral-large-4/)
- [Times of AI: Amodei’s open-weights position (July 27, 2026)](https://timesof.ai/2026/07/anthropic-position-open-weights-models)
- [Times of AI: Anthropic Frontier Red Team on self-replicating IoT risk (October 13, 2026)](https://timesof.ai/2026/10/anthropic-frontier-red-team-self-replicating-ai-iot-devices)
- [Times of AI: Mistral Large 4 preview (October 6, 2026)](https://timesof.ai/2026/10/mistral-large-4-preview-1t-open-weights)