DescriptionRecent industry surveys show that current status quo involving siloed development on proprietary APIs results in significant difficulties in collaborating on AI/ML in the field of EDA. Key issues are: 1) lack of a standard means of exchanging derived data between partners and 2) the concomitant duplication of efforts due to proprietary hurdles. In this special session, experts representing chip design, EDA vendors, and academia provide perspectives on the need for a common API to present derived data for AI/ML pertaining to IC design. This session includes a discussion of existing data formats, applications that use these data types, and of how advances in AI/ML are driving needs which extend beyond capabilities of existing APIs. The topic of IP protection will be a particular focus, including perspectives on what mechanisms need to be in place for data sharing, and various levels of security required depending on the application domain.