与负责任的人工智能实现跨越（英）-埃森哲

From compliance to confidenceEmbracing a new mindset to advance responsible AI maturityPage 03Executive summary: rethinking responsible AIPage 12Becoming reinvention ready: the milestones of responsible AI maturityPage 05 Risky business: the three challenges of the current risk landscapePage 15Responsibility reality check: how ready are companies for responsible AI?Page 10 From compliance to value: companies are acknowledging the impact of responsible AIPage 22Ready, set, grow: five priorities for responsible AIContentsResponsible AI: From compliance to confidence2Executive summary: rethinking responsible AIAs generative AI becomes pervasive in business and society, so too do the risks that come with using it. Consider the chatbot that gave incorrect advice to a customer, creating liability for the company that launched the chatbot. Or the employee who accidentally exposed proprietary company data after using ChatGPT. Or the algorithm that wrongly flagged thousands of individuals for fraud.In this new environment, the responsible development and use of data and AI becomes a critical enabler for organizations to both minimize the risks that come with using these technologies and to unlock its many benefits—from helping individuals perform tasks better and faster to allowing companies to reinvent themselves to create more value and gain a competitive edge through innovation. This is, of course, easier said than done, for the challenges of scaling responsible AI across an organization are daunting. AI-related risks continue to accumulate as generative AI creates and accelerates both data and AI risks. New AI-focused laws and regulations are germinating everywhere. And AI value chains1 are growing in complexity, especially as more companies become both developers and buyers of AI models. Responsible AI 101What is responsible AI?Responsible AI implies taking intentional actions to design, deploy and use AI to create value and build trust by protecting against the potential risks of AI. What is mature responsible AI? Having fully operationalized responsible AI efforts as a platform to take a more systemic, future-orientated approach that unlocks the true value of AI. The North Star for responsible AI maturity is to become a pioneer. No companies are yet at this stage.How do you build a responsible AI program?Our research and work advising clients has shown that all companies can benefit from focusing on these five priorities to improve their maturity and begin to reap the benefits of AI.01Establish AI governance and principles02Conduct AI risk assessments03Systemic enablement for responsible AI testing04Ongoing monitoring and compliance 05Workforce impact, sustainability, privacy, securityResponsible AI: From compliance to confidence3Today’s era of generative AI creates new requirements when it comes to responsible AI. To be a leader in responsible AI, companies must pursue an anticipatory mindset, commit to continuous improvement and extend their