Turstworhty AI

We follow the Ethics Guidelines for Trustworthy AI set up by the European Union for trustworthy AI that is transparent, validated, and free of discrimination. We create ethical recommendation systems and help the validation and assessment of third-party systems to avoid undetected biases. We believe that AI-driven sales bots should comply with competition and consumer protection rules just like the human salesforce must do.

  1. Data integration: Our experience shows that of an organization’s investment into AI, 80% typically goes toward data processing and validation, another significant portion to external data acquisition, and only a very small percentage toward training algorithms. We help small organizations to pool datasets of sufficient size and at a low enough cost to be able to deploy their own AI systems or to devote resources to detecting biases that work against their goals in large corporate systems. See our Music Observatory demo project and our other services to find the relevant, high quality, and affordable data you need to start training algorithms, or to test the effects of existing algorithms on your business or policies.

Related on our Data&Lyrics blog: Why Did We Start The Demo Music ObservatoryFeasibility Study For The Establishment Of A European Music Observatory & The Demo Observatory.

  1. Ethical recommendation systems: We help our clients to power educational, sales, and evaluation applications based on ethical AI that is free of hidden gender, racial, ethnic, and other prohibited biases. See our Listen Local demo project.
    Related on our Data&Lyrics blog: Listen Local: Why We Need Alternative Recommendation SystemsThe Racist Music Algorithm.

  2. Validation: We help our clients to understand what kinds of illegal biases may be affecting their sales and promotion processes when using black-box, third party algorithms on various social media, streaming, and sales platforms. Related on our Data&Lyrics blog: Who Is Not Recommended On SpotifyForgetify: Popular Music That Nobody Listens To.

Daniel Antal
Daniel Antal

My research interests include reproducible social science, economics and finance.