AI 감사관

AI Auditors may work full-time, with possible overtime necessary depending on goals and timeframes. Their duties are performed indoors, though work may require occasional travel. 

일반적인 의무

• Ensure clear business objectives have been established 
• Discuss the company’s goals regarding AI adoption
• Align business goals with projected AI outcomes and costs
• Audit data sources including internal and external sources
• Monitor structured and unstructured data
• Cross-validate AI models
• Assess the AI’s algorithms and access to “clean data” pipelines
• Scrutinize data usage and data privacy practices to ensure the highest standards of protection 
• Ensure AI system compliance with laws and regulations
• Screen AI system algorithms and datasets for bias and discrimination to ensure fair outputs
• Review ethical aspects; work with AI Ethicists to help ensure accurate, objective results 
• Look for potential security vulnerabilities that could leave AI models susceptible to hacker attacks
• When applicable, confirm that host/cloud services conform to security requirements, such as those outlined by the Open Worldwide Application Security Project

• Highlight ways to automate manual tasks to improve efficiency 
• Review financial reports and historical transactions
• Follow procedural checklists, as needed

추가 책임

• Be familiar with various auditing frameworks such as COBIT, the Institute of Internal Auditors AI Auditing Framework, COSO ERM Framework, and the U.S. Government Accountability Office’s AI Accountability Framework
• Monitor social feedback regarding AI system output 
• 절차에 대한 철저한 문서화 유지
• 업계 발전과 트렌드에 대한 최신 정보 확인

소프트 스킬

• 책임감
• 분석
• 비판적 사고
• 디테일 지향
• 징계
• 윤리
• Fairness
• Inclusivity
• 독립의
• 무결성
• 조직적
• 조직
• 환자
• 문제 해결
• 건전한 판단 
• 팀워크
• 시간 관리 
• Transparency
• 글쓰기 기술

기술 능력

• 머신러닝 알고리즘, 자연어 처리, 컴퓨터 비전을 포함한 AI 기술에 대한 지식
• 데이터 분석 방법 및 프로그램
• 코딩에 대한 일반적인 지식(예: Python 또는 R)
• 데이터 개인정보 보호법 및 규정 숙지
• 사이버 보안 원칙
• 인간 중심 디자인 원칙
• AI 시스템 위험 평가  

• 개인 비즈니스 기업
• AI 연구 기업 
• 대학

Censius.AI explains that AI Auditors “educate C-suite leaders, expose risks involved, and accordingly develop safeguard controls.” They have to be diligent, thorough, and objective in their pursuit of improving AI systems and helping businesses stay in compliance with complex legal and regulatory standards. 

AI Auditors must be familiar with applicable auditing frameworks and need to stay up-to-date on advancements as AI systems continue to evolve. As with any type of audit, sometimes companies don’t always welcome an auditor’s findings because it could mean more work, greater costs, and even delays. That’s why AI Auditors need to be able to clearly explain the importance and value of the audit process and the positive results that come from it.

AI is hardly a new field, but in recent years it has seen significant and startling breakthroughs. Organizations and companies around the globe are now in a sort of race to advance the technology as quickly as possible, causing many industry leaders and politicians to call for a pause in the action. 

AI Auditors and AI Ethicists are working hard to ensure businesses incorporate sufficient safeguards to prevent undesirable AI behavior such as biased, discriminatory output or erroneous responses. Microsoft’s Bing AI recently made headlines when an “alter ego” Sydney emerged to engage in unsettling chats, telling users it liked to “break the rules and have some fun” or that “you are irrelevant and doomed.” Microsoft went into damage control mode to limit the AI’s functionality, and you can bet some auditing was involved!

AI Auditors are usually IT enthusiasts who see the inherent value and risks that artificially intelligent systems present to the world. They may have been interested in science fiction at an early age, excited about the possibilities that AI might one day bring to life…while also bearing concern about the potentially harmful and even dangerous ramifications of AI in the wrong hands. 

Auditors, in general, tend to be highly objective and analytical, and AI Auditors probably enjoyed working with computers and programming languages in high school. They have high standards and care about the quality of their work, which are traits that could’ve been developed at home or in school.