Context: The US-UK agreement represents a major step towards international cooperation in ensuring the safe and responsible development of AI technologies. It highlights the shared commitment of both nations to address the challenges and risks associated with advanced AI systems.
The agreement reflects the increasing global awareness of the potential risks and benefits of AI and the necessity for collaborative efforts to guide its development. It underscores the importance of establishing international standards and guidelines for AI safety, security, and ethics.
AI Safety and Security
- US-UK Agreement
- The agreement facilitates the sharing of critical information on AI capabilities, risks, and best practices between the US and the UK.
- It promotes the alignment of approaches to ensure the safe deployment of AI systems and enables joint testing exercises to assess the performance and reliability of AI models.
- The agreement also encourages personnel exchanges between AI Safety Institutes to foster collaboration and knowledge sharing.
- Follows commitments made at the Bletchley Park AI Safety Summit (2023)
- Importance of AI Safety
- AI safety measures are crucial to address the potential risks posed by advanced AI systems, such as algorithmic bias, privacy violations, and security vulnerabilities.
- Ensuring the safety of AI systems is essential to prevent unintended consequences and protect individual rights and societal values.
- Responsible AI development involves transparency, accountability, and the incorporation of human oversight and control.
- Potential Impact of the Agreement
- The US-UK agreement sets the stage for enhanced global cooperation in AI safety and security, encouraging other nations to follow suit and collaborate on this critical issue.
- The agreement serves as a model for other nations to emulate, inspiring them to forge similar partnerships and prioritize AI safety in their own AI development efforts.
- The agreement recognizes the potential risks of AI in spreading misinformation and undermining election integrity, and seeks to develop strategies to counter these threats.
AI Regulation and Policy
- US Efforts
- The National Telecommunications and Information Administration (NTIA) in the US has initiated a consultation process to gather insights on the risks, benefits, and potential policy implications of open-source AI models and dual-use foundation models.
- President Biden issued an executive order in 2023 that outlines the US government's commitment to ensuring the safe and responsible deployment of AI systems.
- In 2022, the White House released a Blueprint for an AI Bill of Rights, which sets forth principles and guidelines for protecting individual rights and promoting the responsible use of AI.
- EU AI Act
- The proposed European Union AI Act seeks to establish comprehensive safeguards on the use of AI systems, with specific provisions for high-risk applications such as law enforcement.
- The act aims to ensure that AI systems are transparent, explainable, and subject to human oversight, while also empowering consumers to challenge decisions made by AI systems.
- The EU AI Act recognizes the potential for AI misuse and seeks to establish clear accountability mechanisms for AI developers and deployers.
- India's Approach
- India's Ministry of Electronics and Information Technology (MeitY) has issued evolving advisories on the deployment of AI systems in the country.
- The advisory, issued on March, 2024, directed intermediaries to label any under-trial/unreliable artificial intelligence (AI) models, and to secure explicit prior approval from the government before deploying such models in India.
- The Indian Government is developing an AI regulation framework, set for release in mid-2024, with the intention of harnessing AI for economic growth and addressing potential risks and harms.
Open-Source AI Models and Implications
- Prominent Examples
- Meta has released Code Llama 70B, the largest and best-performing model in the Code Llama family. Code Llama is a state-of-the-art large language model (LLM) capable of generating code, and natural language about code, from both code and natural language prompts.
- OpenAI's ChatGPT has been released through a controlled API and product-based approach.
- Dual-Use Foundation Models with widely available weights, enabling both beneficial and malicious applications
- Implications for Innovation and Competition
- In 2024, open-source pretrained AI models have gained significant traction, empowering businesses to accelerate growth by combining these models with private or real-time data.
- Generative AI challenges a core tenet of traditional intellectual property frameworks: only works created by humans are protected by copyright laws.
- Emerging use cases around generative AI are disrupting traditional views of creativity, authorship, and ownership and pushing the boundaries of copyright law.
- In 2024, open-source technology faces increased scrutiny as its prolific use, including in proprietary coding, raises the need for pervasive security screening.
Implications for India
- The AI advisory in India emphasizes transparency, content moderation, consent mechanisms, and deepfake identification to ensure responsible AI deployment and safeguard electoral integrity.
- AI presents significant opportunities for economic growth in India. The AI industry is estimated to grow year-over-year at a CAGR of 30%. India's AI market is growing at a CAGR of 25-35% and is projected to reach around $17 billion by 2027.
- However, the adoption of AI technologies may lead to job displacement in certain sectors. As per market trends, more than 16 million working employees in India will need reskilling and upskilling due to AI's influence by 2027.
- AI technologies have the potential to enhance law enforcement capabilities in India. However, the use of AI in law enforcement also poses risks, such as bias, privacy violations, and potential misuse of power.
- An updated toolkit for responsible AI practices in law enforcement has been released by INTERPOL and UNICRI in 2024.
- The NITI Aayog released an approach paper that explores the various ethical considerations of deploying AI solutions in India.
Opportunities for India
- The digital divide in India is being addressed. There were 751.5 million internet users in India at the start of 2024, when internet penetration stood at 52.4 percent. Initiatives like BharatNet aim to bridge the digital divide and potentially lead to a major positive shift.
- The Cabinet has approved the comprehensive national-level IndiaAI mission with a budget outlay of Rs.10,371.92 crore. The IndiaAI mission will establish a comprehensive ecosystem catalyzing AI innovation through strategic programs and partnerships across the public and private sectors.
- India holds a prominent global position in AI skill penetration and talent concentration, showcasing a strong base of AI professionals. There were 4.16 lakh AI professionals, poised to meet the increasing demand expected to reach 1 million by 2026.
- AI-driven platforms deliver insights to farmers on topics like disease risks, yield forecasts, labor needs, crop protection, weather impacts on crops, and harvest windows.
- AI has been used thoughtfully by educators to support learning and to give them "time back" in their day. AI applications in education will be overwhelmingly administrative.
Way Forward for India
- Developing a National AI Strategy
- Establishing a dedicated AI governance framework and regulatory body
- Allocating resources and creating incentives for AI research and innovation
- The Prime Minister, Shri Narendra Modi inaugurated the Global Partnership on Artificial Intelligence (GPAI) Summit. GPAI is a multi-stakeholder initiative with 29 member countries aiming to bridge the gap between theory and practice on AI.
- India is the lead chair of GPAI in 2024.
Countries like Japan, Rwanda, Benin, Egypt, Morocco, Mauritius, Tunisia, Sierra Leone, and Senegal have developed comprehensive AI strategies and governance frameworks. The Hiroshima AI Process was launched by the G7 under Japan's presidency in May 2023, with the aim of promoting safe, secure, and trustworthy AI.