UN’s First Scientific Report on AI

Syllabus: GS3/Science and Technology

In News

• The first global scientific assessment on AI has been released by the United Nations convened Independent International Scientific Panel.

Major Findings 

• AI development outpaces regulation: AI capabilities, especially frontier AI models and autonomous AI agents, are advancing faster than scientific understanding and existing governance mechanisms, increasing the risk of serious harm.
• Weak and fragmented AI governance: Current AI regulations and ethical frameworks are fragmented, dominated by a few corporations, and lack effective methods to independently assess AI capabilities and risks.
• Growing global computer divide: AI leadership increasingly depends on access to advanced computing infrastructure, semiconductors, cloud facilities, and datasets.
  • The United States accounts for around 75% of global AI computing capacity and China another 15%, leaving the rest of the world to share the remaining 10%.
  • This “compute divide” risks making many countries AI consumers instead of AI creators.
• Risks of AI concentration: Few companies and countries are increasingly controlling the AI value chain from advanced chips and cloud infrastructure to foundation models and deployment platforms driving a concentration of AI development. 
• Democratic Threats: AI deepfakes and persuasion algorithms are eroding shared truth, creating disinformation and impacting democratic processes. 
• Geopolitical concerns: Dependence on a few AI powers could undermine countries’ strategic autonomy in critical sectors such as healthcare, defence, public services, and infrastructure.
• Implications for India: The findings reinforce the importance of initiatives such as the IndiaAI Mission and investments in domestic AI computing infrastructure to reduce long-term technological dependence.

Recommendations

• It urges governments to improve governance and build domestic AI capabilities (Sovereign AI Models), and to work together internationally to ensure AI is safe, inclusive and beneficial for all.
• There is a need for better and more harmonised global AI governance.
  • Independent scientific assessments of risks from AI.
  • Better access to scientific resources and computing infrastructure.
  • Policies that promote competition and open scientific research.
  • Investing in public-interest AI and expanding the participation of developing countries.
  • Apply established human rights and child rights impact assessments.

Source :IE