【日本】Four Steps to Implement Responsible AI in Your Organization: How to Integrate It into Daily Operations and Make It Stick

With the approval of the EU AI Act, establishing AI ethics and integrating responsible AI into organizations has become an urgent issue for companies. However, putting this into practice is far from easy. This article explains four key actions (translation, integration, alignment, and dissemination) that leaders should take to integrate responsible AI practices into broader operational standards, as identified by the authors.

When the EU AI Act was approved by the European Parliament in March 2024, Deutsche Telekom felt confident and prepared. Since establishing its Principles for Responsible AI in 2018, the company has worked to integrate these principles into the development cycle of its AI-based products and services.

Unfortunately, as our research suggests, such proactive measures are the exception, not the norm. While AI ethics is a critical issue for many organizations, translating AI principles into practice and action is easier said than done. However, considering the severe fines for non-compliance, the situation is urgent. What should leaders do to strengthen their responsible AI initiatives?

To find the answer, we visited organizations across various industries and discovered that there are distinct stages in the journey toward adopting responsible AI. While data engineers and data scientists typically bear most of the responsibility for the AI development lifecycle from conception to production, non-technical leaders can play a crucial role in ensuring the integration of responsible AI. We identified four key actions leaders should take to fully integrate responsible AI practices into broader operational standards: translation, integration, alignment, and dissemination.

Many organizations have developed AI ethics charters but often struggle to introduce AI principles into daily operations.

indicating the need to translate these principles into practical guidance.