Many newsrooms rely on a human-in-the-loop approach when developing AI tools. But is human oversight alone enough to address the risks involved in building AI-powered products and workflows?
In this webinar, Alessandro Alviani from Süddeutsche Zeitung, explores why the traditional human-in-the-loop paradigm is insufficient and why newsrooms need to think more broadly about risk management and human involvement throughout the AI development process.
Drawing on concrete examples from Süddeutsche Zeitung, including the Wirecard voice assistant and election assistants, the session will focus on three key insights:
- There are no risk-free AI products.
- How we design for risk minimisation and mitigation matters.
- Humans play a crucial role throughout the development process – even when they are not directly checking the answers generated by an AI chatbot.
Through these examples, the webinar will examine how human oversight can be understood beyond simply reviewing AI-generated outputs.
About the presenter:
Alessandro Alviani is Lead for Generative AI at Süddeutsche Zeitung Digitale Medien, the digital publisher of the German newspaper Süddeutsche Zeitung. He has a background in journalism and product development, and is experienced in building and evaluating AI-driven products, tools, and workflows for readers, editors, and business stakeholders.
His work is increasingly focused on questions around human-AI interaction and oversight of AI systems. Previously, he was Product Lead AI at Ippen Digital and Editorial Director at Microsoft News in Berlin. Before that, he covered Germany for the Italian daily La Stampa. He is based in Munich.