Companies Committed to Responsible AI: From Principles towards Implementation and Regulation?

Area of Science:

• Computer Science
• Ethics in Technology
• AI Governance

Background:

• The concept of responsible AI emphasizes fairness, transparency, security, privacy, accountability, and societal benefit.
• Numerous organizations, including 24 AI companies, have publicly committed to these principles since 2016.
• Commitments were made via website pledges and participation in the 'Partnership on AI'.

Purpose of the Study:

• To investigate the practical implementation of responsible AI principles by signatory companies.
• To understand companies' perspectives on the role of external societal actors, particularly government, in AI regulation.
• To assess whether companies are actively integrating ethical AI practices or engaging in 'ethics washing'.

Main Methods:

• A comprehensive web search was conducted to gather data.
• Analysis focused on the actions and public statements of 24 AI companies that pledged to responsible AI principles.
• The study examined implementation efforts and views on AI regulation.

Main Results:

• Approximately three major AI companies have made significant strides in implementing responsible AI, notably by developing and open-sourcing software tools.
• These leading companies appear to be genuinely committed, avoiding 'ethics washing'.
• Around ten US and European companies support governmental regulation for AI, particularly for high-risk applications.

Conclusions:

• Leading AI firms are actively developing tools to operationalize responsible AI principles.
• A consensus is emerging among some companies for a balanced regulatory approach, focusing on high-risk AI.
• The study highlights a move towards practical implementation and a call for thoughtful AI governance.