Responsible AI implementation in the generative age
- Validity and reliability
- Safety
- Security and resiliency
- Accountability and transparency
- Explainability and interpretability
- Privacy
- Fairness with mitigation of harmful bias
To investigate the current landscape of responsible AI across the enterprise, MIT Technology Review Insights surveyed 250 business leaders about how they’re implementing principles that ensure AI trustworthiness. The poll found that responsible AI is important to executives, with 87% of respondents rating it a high or medium priority for their organization.
A majority of respondents (76%) also say that responsible AI is a high or medium priority specifically for creating a competitive advantage. Few have managed to make these ideas a reality. Only 15% of respondents felt prepared to adopt responsible AI practices despite their importance.
At the age of generative AI, putting responsible AI into action requires adopting a set of best practices. Leading companies are doing this. Cataloging AI models, data, and implementing governance control are some of these practices. The companies may benefit by conducting thorough assessments, tests, and audits to ensure compliance with regulatory requirements, security, or risk. To ensure that their change efforts are successful, companies should empower their employees by providing training in a large scale. “We know AI is the biggest change in technology we’ve ever seen, but it’s not being implemented properly,” says Steven Hall. He’s chief AI officer at ISG and president of EMEA. The report is available for download. This content was produced by Insights, the custom content arm of MIT Technology Review.
