Context: The Comptroller and Auditor General of India (CAG), who is the chair for the Supreme Audit Institutions (SAIs) of the G20, warned of absolute dependence on Artificial Intelligence (AI).
Artificial Intelligence (AI):
- In Layman terms, Artificial Intelligence can be defined as a branch of science that makes machines or software to mimic human intelligence such as recognizing speech, making decisions, and identifying patterns.
- AI is an umbrella term that encompasses a range of technologies, including machine learning, deep learning, and natural language processing (NLP).
Examples of Artificial Intelligence:
- ChatGPT: Uses large language models (LLMs) to generate text in response to questions or comments posed to it.
- Google Translate: Uses deep learning models to translate text from one language to another.
- Netflix: Uses machine learning to create personalized recommendations for users based on their previous viewing history.
- Tesla: Uses computer vision to power self-driving features on their cars.
Role of Artificial Intelligence in Governance:
1. Smart Policy Making:
- Data-oriented Policy making: Collection and analysis of data such as demographic, behavioural data etc. from different ministries and government sectors to effectively formulate the policy and schemes.
- Localisation of Policy Making: AI-enabled tools have the potential to provide regional and local leaders with insights and analysis allowing policies to be better tailored to local conditions.
- Timely Awareness Generation: Use of language models to create awareness in different languages in real-time using LLM can increase the outreach of government initiatives.
- Measurement of effectiveness: The government can create AI-backed feedback mechanisms to measure satisfaction of different sections from different platforms based on different parameters.
- Futuristic Policy Making: AI through data provided can predict emerging issues and can improve preparation to tackle such issues.
2. Service Delivery:
- Effectiveness of Service Delivery: AI systems can reduce the time, cost and increase the effectiveness of the services provided by the governments. E.g.,
- AI-enabled procurement processes can allow governments to identify inefficiencies and potential cost savings in the products and services they purchase.
- Traffic management through AI-based modelling can significantly reduce the amount of time spent in traffic.
- Allocating health system resources using AI-enabled patient demand analytics can minimize wait times while reducing costs.
- By using job seeker data and their skill-set data in AI-enabled systems government can plan job schemes.
- Automation of Delivery through AI: Through AI-enabled system, the government can create a single point entry for all citizens and can update it regularly. E.g., if a salaried person gets unemployed her unemployment allowances and other scheme benefits start automatically through an AI-backed system.
3. Efficient operations:
- Learning and Training: AI can help enhance the learning and development of organizations and employees through customized training and education programs.
- Rationalisation of Work: AI can help in decreasing the repetitive tasks handled by government officials.
- Effective Auditing: AI can help in the allocation, expenditure and monitoring of the finances for initiatives and give effective guidance for the optimisation of financial and human resources.
Concerns associated with the use of AI in Governance:
- Propagate bias: AI systems can perpetuate and even amplify biases in decision-making. Algorithms through machine learning if trained on biased data can impact policymaking.
- Privacy concerns: AI systems can collect and process large amounts of personal data which raises concerns about privacy and data security demanding effective legal safety for its large-scale use in government functioning.
- Risk of exclusion and discrimination: Algorithms of AI can be so complex that even those who created the algorithm cannot thoroughly explain how the variables led to the resulting prediction. This opacity poses the risk of exclusion and discrimination by AI in decision-making.
- Loss of jobs: AI systems have the potential to automate tasks traditionally performed by humans. E.g., Chatbot replaced human-assisted consumer support, which could lead to job displacement and cause socioeconomic disruption.
- Security risks: AI can be used for malicious purposes halting the regular functioning of government e.g., cyberattacks, deep fake etc.
- Ethical concerns: There is no legal accountability and responsibility arising out of the decisions made by AI.
- Political Concerns: Data collected by the government if used by AI to predict voter behaviour and its manipulation during election times threatens the very foundation of democracy.
Challenges faced in auditing using AI:
- Data standardisation: Lack of authorised sources and challenges faced in ensuring the accuracy of vast data.
- Data Integration: Lack of synchronisation, data integration and cross-referencing among different departments and ministries.
- Lack of Capacity: Lack of capacity and awareness about AI among auditors.
- Lack of AI regulation: Lack of explicit National guidelines and regulations for the use of AI in audits in India.
- Comprehensive regulatory framework: India needs a comprehensive regulatory framework with both horizontal regulations (that would be applicable across sectors) and vertical regulations (sector-specific). There is a need to identify the capabilities of AI that are more susceptible to misuse than others to promote the responsible use of AI.
- Responsible use of AI: The AI system should be programmed to maintain the integrity of data and should work on the principle of justice and promoting individual liberty and privacy. Use authentic data sources to ensure transparency, address legal concerns, and look at deficiencies in IT controls and governance.
- European Union’s Artificial Intelligence Act:
- The Act was formulated:
- To bring transparency, trust, and accountability to AI.
- To create a framework to mitigate risks to the safety, health, fundamental rights, and democratic values of the EU.
- To address the different levels of risks E.g., limited risk, high risk and unlimited risk as defined in the act.
- Under the EU’s General Data Protection Regulation (GDPR), Data Protection Impact Assessments are legally required if organisations use AI systems that process personal data to avoid potential risks.
- The Act was formulated:
- The U.K.’s Information Commissioner’s Office has published draft guidance on the AI auditing framework.