#ainow — Public Fediverse posts

undefined | Concerns about Mythos lie in its ‘purposely vague language’ – AI expert by Krishnan Guru-Murthy

**Concerns about Anthropic’s new AI model, Mythos, have been highlighted by AI experts. In a recent Channel 4 News segment, presenter Krishnan Guru‑Murthy sat down with Dr Heidy Khlaaf, chief Artificial Intelligence scientist at the AI Now Institute, to discuss the implications of the model’s release.**

**Dr Khlaaf warned that the biggest problem with Mythos lies in its “purposely vague language,” which makes it difficult to assess how the system will behave in real‑world situations. She explained that this lack of clarity obstructs transparency, hampers accountability, and could enable the model to be misused without clear safeguards.**

**While acknowledging that the technology is powerful, Khlaaf stressed that the threat it poses depends on how it is deployed. She called for stricter oversight, clearer documentation, and rigorous testing to ensure that Mythos does not become a tool for harmful applications, urging policymakers and developers to address these gaps before the model is widely released.**

Read more: undefined

#ai #technology #anthropic #ainow #mythos

#AI Now Institute Releases “Artificial Power: 2025 Landscape Report”
https://www.infodocket.com/2025/06/03/ai-now-institute-releases-artificial-power-2025-landscape-report/ #AINow

#AINow fordert aus guten Gründen strenge Regulierung für GPAI - das könnte auch passieren, aber läuft ins Leere, wenn mit Foundation Modellen nun neue Kategorien geschaffen werden, die wieder zu unreguliert bleiben. #AIA #AIAct

https://ainowinstitute.org/publication/gpai-is-high-risk-should-not-be-excluded-from-eu-ai-act