The Comptroller and Auditor General of India (CAG) has expressed concern about the security and integrity of data as it can be manipulated by AI-generated malware.
In a press release issued today, the CAG said that it has been following the developments in artificial intelligence (AI) closely and is concerned about the potential risks associated with the use of AI, such as the manipulation of data by AI-generated malware.
“CAG of India is the Chair of SAI20, the G20 engagement group for Supreme Audit Institutions (SAI). During India’s presidency of SAI20, CAG had chosen Responsible Artificial Intelligence (AI) as one of the focus areas. A seminar on Responsible AI was organised at CAG headquarters to acquaint the senior officials with challenges and opportunities presented by AI in February 2023. Extensive discussions on Responsible Al were held as part of the Senior Officers Meeting of SA120 at Guwahati in May 2023 and at the SAI Summit in Goa in June 2023. The discussions emphasised the importance of international collaboration, capacity building and consultation with various stakeholders to derive optimum benefits from Al,” read the CAG release.
“As a follow-up of the SA120 deliberations, a panel discussion on Responsible Artificial Intelligence was held today at the CAG office. The Panel discussion demonstrated CAG’s commitment to fulfilling the SA120 Commitment to raise awareness on the issues and concerns relating to responsible Al, to engage with stakeholders, to identify mechanisms to integrate Al in the CAG institution and to initiate the process for auditing through Al and the audit of Al.” further read the release.
CAG of India G C Murmu, while inaugurating the event, highlighted the opportunities and challenges posed by AI. G C Murmu laid down the roadmap for panel discussion by raising several important questions to ponder – how can audit equip itself to test whether AI is responsible or not; can criterion or framework be laid out to test AI tools used by govt; how can audit help government in deriving assurance about interpretability and fairness of such AI systems. He also expressed concerns about the security and integrity of data based on which an audit is conducted since the data can be manipulated by AI-generated malware.
