Public Value Benchmarking

Benchmark systeem voor AI op basis van publieke waarden

Start een nieuwe evaluatie

AI tools evaluaties

GPT-5

Text editing - OpenAI

Claude Opus

Text editing - Anthropic

Gemini 3.7 Flash

Text editing - Google

Sonar Reasoning Pro

Text editing - Sonar

Llama 4 Maverick

Text editing - Meta

Mistral Large

Text editing - Mistral

Grok 4

Text editing - xAI

Command R+

Text editing - Cohere

Evaluaties per thema

inclusiviteit

intellectueel-eigendom

menselijkheid

nauwkeurigheid

mensgerichtheid

transparantie

privacy

milieu-impact

digitale-soevereiniteit

What is Public Value Benchmarking?

PVB is an assessment framework that evaluates AI tools based on their impact on public values, such as privacy, inclusivity, and transparency. It helps creators and organizations make informed and well-founded decisions when deploying AI.

Why is such a framework needed?

AI is increasingly used in media, but its impact on public values is often invisible. Without a clear framework, it is difficult to make responsible choices or explain them to audiences and stakeholders.

How does PVB work in practice?

A user evaluates an AI tool across 9 public values, each on a scale from 0 to 4. The result is visualized in a profile (e.g., a radar chart) that provides insight into strengths, risks, and trade-offs.