Mark Schwartz on AI ethics
From Software Designing & Software Development Is An Art In Itself | Mark Schwartz | AWS | AIM TV · · AIM Network
“Ethical is not the same as compliant — it's very easy to confuse those two things; the rulebook doesn't really exist today for generative AI, so you need a different approach to thinking ethically.”
On , Mark Schwartz, Senior Vice President of AECO Software at Trimble Inc., spoke about AI ethics during Software Designing & Software Development Is An Art In Itself | Mark Schwartz | AWS | AIM TV on AIM Network.
About Mark Schwartz
Mark Schwartz, Senior Vice President of AECO Software at Trimble, has discussed his experiences as a former CIO for U.S. Citizenship and Immigration Services (USCIS) and his work at Amazon Web Services (AWS). In a December 2024 interview, Schwartz said he is not yet convinced that AI can perceive beauty in code the way a human can, though he acknowledged AI can write code and perform tasks such as writing automated tests and finding defects. He also stated that ethical considerations for generative AI are distinct from compliance, as no rulebook currently exists for the technology. Schwartz described a project where AWS worked with the International Centre for Missing and Exploited Children to use image recognition to search for missing children, and another effort to help homeless people access housing by providing a way to store documents securely online. In a 2020 presentation, Schwartz recounted his efforts to transform USCIS's IT operations, moving from an average release cycle of 18 months to deploying code multiple times a day. He described navigating bureaucratic processes, such as a policy requiring 87 documents and 11 gate reviews, by creating a tailored process that effectively reversed the original requirements. Schwartz argued that bureaucracy can be tackled and used to drive transformation, noting that DevOps automates rules to enforce good practices, similar to traditional bureaucratic controls but with greater efficiency. He emphasized that bureaucracy does not have to be wasteful or unchanging, and that organizations can separate desired outcomes from best practices to allow for continuous learning.