Shared recording

AI, Young India and Future Opportunity Panel

Session 12.mp4 · 1:08:17 · 5 speakers

Download (4.2 GB) Anyone with this link can watch and download this recording.

Key takeaways

  1. AI was framed as both a productivity tool and a potential means to address national-scale challenges in education, healthcare, governance, and frontline services.
  2. The panel distinguished basic access to AI tools from meaningful access, which also requires skills, institutional support, language accessibility, and inclusion.
  3. Young people expressed concern that AI may reduce entry-level opportunities and raise employers’ expectations of new entrants.
  4. Speaker 4 described AI’s employment effect as a shift in roles and skills, rather than straightforward replacement of workers.
  5. Speaker 3 argued that AI lowers barriers to building products, potentially enabling more entrepreneurs and non-technical builders.
  6. Concerns about bias, data annotation, marginalised communities, privacy, sustainability, and human cognition remained open and require continued attention.
  7. No formal decisions or assigned actions were recorded.

Key moments

  • Session opening and panel introduction
    Speaker 1 frames the discussion around AI, young Indians, skills, access, work, and opportunity.
  • AI as an assistant and tool for inclusive creation
    Speaker 3 argues that AI should enhance creativity and problem-solving, including for underserved communities.
  • Youth perspective on AI power and inclusion
    Speaker 1 broadens the discussion beyond LLMs to hiring, public services, data annotation, and algorithmic control.
  • Employment and entry-level work discussion
    Speaker 1 outlines young people’s concerns about AI replacing repetitive and entry-level work.
  • Enterprise response on job shifts
    Speaker 4 says SAP’s hiring and training continue, but skills requirements have shifted toward AI and cloud.
  • Education-system critique
    Speaker 5 argues that education quality, accountability, and investment must improve.
  • AI risks, adoption, and safety
    Speaker 3 discusses entrepreneurship, adoption gaps, regulation, safety, and government applications.
  • Audience Q&A on cognition and sustainability
    An audience member raises concerns about human agency, cognition, environment, education, and Gen Z insecurity.
  • Question on AI investment bubble
    An audience member asks who would bear the cost if AI investment assumptions fail.
  • Challenge on AI externalities
    Speaker 1 challenges whether AI can avoid displacement and other harms; Speaker 3 responds on proactive risk management.
  • Closing remarks
    Speaker 2 ends the session and Speaker 1 thanks participants, describing the discussion as a starting point.

Decisions made

No explicit final decision was identified.

Proposals & suggestions (not confirmed decisions)

  • Use AI as a creative, critical-thinking assistant or teammate rather than simply a mechanism to obtain answers. — suggested by Speaker 3. Outcome: Discussed as Speaker 3’s recommended approach; no formal decision.
  • Treat application of AI to real-world problems, including underserved communities and national-scale services, as a valid form of AI creation. — suggested by Speaker 3. Outcome: Supported by examples from Speaker 3; no formal decision.
  • Invest in understanding AI skilling potential for young people. — proposed by Speaker 1. Outcome: Raised as a need; no formal decision.
  • Expand AI access through language support, Tier 2 and Tier 3 outreach, and small innovation centres. — proposed by Speaker 4. Outcome: Presented as areas for institutional support; no formal decision.
  • Include people affected by AI systems in AI co-creation, including marginalised communities and workers involved in data annotation. — proposed by Speaker 1. Outcome: Discussed; Speaker 3 stated that engagement with civil society and underserved communities is occurring, but no formal decision.
  • Use AI in government for citizen-service delivery and internal efficiency, including multilingual, interoperable services. — proposed by Speaker 3. Outcome: Presented as an opportunity; no formal decision.
  • Develop regulatory frameworks and AI safety institutions more rapidly to keep pace with model development. — proposed by Speaker 3. Outcome: Presented as a need; no formal decision.
  • Improve education quality, accountability, and investment in both public and private higher education. — proposed by Speaker 5. Outcome: Presented as a personal view; no formal decision.
  • Consider how AI and agentic commerce can extend advantages to MSMEs. — suggested by Speaker 5. Outcome: Raised as a closing perspective; no final response or decision.

Topics discussed

Opening, panel introduction, and session framing

to 00:04:37 Speaker 1, Speaker 2, Speaker 3

Speaker 1 introduced the session, its theme, participating speakers, and the focus on AI’s effect on skills, jobs, access, entrepreneurship, and opportunity for young Indians. Speaker 2 began moderation by asking how India can move from AI adoption to creation and democratise opportunity.

AI adoption, creation, and inclusive opportunity

to 00:27:53 Speaker 1, Speaker 2, Speaker 3, Speaker 4

Speaker 3 described AI as a creative assistant and argued that solving problems with AI is also a form of creation. Speaker 1 raised unequal access and skills concerns. Speaker 4 discussed business AI, language accessibility, enterprise adoption, and innovation support. Speaker 1 called for co-creation with communities affected by AI, while Speaker 3 responded that OpenAI’s stated mission is to ensure AGI benefits all humanity and pointed to AI’s multimodal and language capabilities.

Jobs, entry-level work, and skills

to 00:37:30 Speaker 1, Speaker 2, Speaker 4

Speaker 1 described young people’s anxiety about automation of repetitive, manual, entry-level, data, and creative work. Speaker 4 said SAP continues to hire but has changed training content toward AI and cloud skills. Speaker 1 described peer-led AI learning and argued that humanities-based critical thinking may remain valuable.

Education and skilling system

to 00:43:08 Speaker 2, Speaker 5

Speaker 5 criticised India’s education system as inadequate and called for quality, accountability, and increased investment. Speaker 2 asked where India should begin its reskilling effort.

AI risks, adoption, safety, and government use

to 00:50:42 Speaker 2, Speaker 3

Speaker 3 discussed AI-enabled entrepreneurship, evolving career pathways, current evidence on employment, adoption gaps, safety work, and the need for rapid regulatory and institutional frameworks. Speaker 3 also described possible AI use in citizen-service delivery and government efficiency.

Audience questions: cognition, sustainability, investment risk, and data security

to 01:05:59 Speaker 1, Speaker 2, Speaker 3, Speaker 4, Speaker 5

Audience members asked about human cognition, AI adoption, environmental impact, the possibility of an AI investment bubble, data sovereignty, cybersecurity, and the allocation of AI’s benefits and costs. Speaker 3 addressed sustainability claims, study mode, safety, data-centre impacts, and stakeholder engagement. Speaker 1 raised AI intimacy, privacy concerns, and critiques of the DPDP law. Speaker 4 responded on private AI investment and public financial exposure. Speaker 5 raised AI’s possible value to MSMEs through agentic commerce.

Closing reflections

to 01:08:16 Speaker 1, Speaker 2

Speaker 2 closed the discussion. Speaker 1 thanked the moderator and panel, highlighted the diversity of views and quality of audience questions, and characterised the discussion as a beginning for further work.

Important questions

  • How can India transition from AI adoption to AI creation and democratise opportunity? — asked by Speaker 2
    Answered Speaker 3 said creation should include applying existing AI models to real-world outcomes, especially for underserved communities and national-scale challenges. Speaker 1 and Speaker 4 added that meaningful access requires skills, institutional support, language accessibility, and broader inclusion.
  • What kinds of jobs are at risk from AI? — asked by Speaker 2
    Answered Speaker 1 said repetitive, manual, entry-level, data, and some creative work are perceived as most at risk. Speaker 4 characterised the effect as a shift toward roles requiring AI and cloud skills rather than straightforward job replacement.
  • How can organisations develop talent if entry-level work declines? — asked by Speaker 2
    Answered Speaker 4 said SAP retains recruitment and training mechanisms but has changed the content toward AI use and cloud knowledge. Speaker 3 added that AI may let individuals build independently and pursue less conventional career paths.
  • Where should India start with reskilling and upskilling? — asked by Speaker 2
    Partially answered Speaker 5 argued that India’s education system needs better quality, accountability, and investment. No specific national implementation plan was agreed.
  • How should AI’s psychological and cognitive effects, human agency, sustainability, and education implications be addressed? — asked by Speaker 3
    Partially answered Speaker 3 referred to OpenAI’s “study mode,” which is intended to guide learners through questions rather than simply provide answers, and said research on real-world cognition is being pursued. Speaker 1 suggested literature on AI intimacy and relationships. Sustainability concerns were discussed but no comprehensive framework was identified.
  • If the AI investment boom proves to be a bubble, who bears the cost and what responsibility does the industry have to young people encouraged to build careers around AI? — asked by Speaker 1
    Partially answered Speaker 4 said current investment is interconnected across chips, models, and data centres, with broad private-sector participation, and stated that public exposure is lower because the funding is private. The broader responsibility of industry to young people was not directly addressed.
  • How can India avoid geopolitical vulnerability, data breaches, and cyberattacks while relying on digital public infrastructure and foreign technology companies? — asked by Speaker 1
    Partially answered Speaker 2 deferred a detailed answer. Speaker 1 raised concerns about the DPDP law and personal-data exposure. No full response on geopolitical vulnerability or cybersecurity was provided.
  • Why cannot AI be developed in a way that avoids externalities such as displacement and environmental impacts? — asked by Speaker 1
    Partially answered Speaker 3 said model developers are proactively working on risks through transparency, red-teaming, civil-society engagement, and human-alignment efforts, while acknowledging they will not get everything right immediately. Speaker 3 also said the referenced TCS data-centre work uses existing government-designated data-centre land.

Important numbers & facts

  • Thirteenth PASI Annual Forum — The session was held at the thirteenth PASI Annual Forum.
  • India 3.0 — The forum theme was stated as “India three point o, engaging for growth in a disrupted world.”
  • second largest country — Speaker 3 said India is OpenAI’s second-largest country, in the context of ChatGPT adoption.
  • 60% of Indian unicorns — Speaker 4 stated that almost 60% of Indian unicorns run on SAP.
  • April 2024 — Speaker 3 said they joined in April 2024 while describing early engagement with civil society.
  • two and a half years — Speaker 3 said they had been in the AI space for two and a half years.
  • 10% of time versus 25% earlier — Speaker 3 said AI may reduce their administrative work from approximately 25% of time to 10%.
  • 100 megawatt data center — Speaker 3 referred to work with TCS to build a 100-megawatt data centre.
  • 6.7% error rate monthly — Speaker 1 attributed a monthly 6.7% error rate to work by Ritika Kera when discussing Aadhaar-related exclusion; the precise underlying measure was not further explained.
  • 2 rupees — Speaker 1 said an Indian Freedom Foundation report stated that personal data was being sold online for as little as 2 rupees.
  • 42,000 students — Speaker 5 said Texas A&M had a sanctioned student strength of 42,000 when Speaker 5 was a student.
  • 56,000 students — Speaker 5 said Ohio State University had 56,000 students at that time.
  • 1.4 billion people — Speaker 5 referred to India as having 1.4 billion people while comparing higher-education capacity.

Speaker highlights

Speaker 1

Main points

  • Introduced the session and framed questions about AI, youth skills, access, jobs, entrepreneurship, and innovation.
  • Argued that AI access is not equal merely because a tool is available; meaningful access depends on skills and institutional conditions.
  • Raised young people’s concern that automation is reducing repetitive, manual, entry-level, data, and some creative work.
  • Called for AI co-creation that includes marginalised groups, data annotators, and communities affected by data-centre development and AI systems.
  • Argued that data and AI outputs are not neutral, and raised privacy concerns including the DPDP law and potential exposure of personal data.
  • Closed by thanking the panel and audience and describing the discussion as a starting point.

Positions

  • AI should not be treated as a purely technical issue; it affects social equity, public services, livelihoods, and participation.
  • Young people should be included as active contributors to AI design and governance, not only as passive recipients.
  • The costs and externalities of AI need attention alongside its promised benefits.

Questions raised

  • Do young people have the skills needed to use AI for innovation, rather than only personal or routine work?
  • What is preventing meaningful AI innovation by young people: access, institutional priorities, skills, or other factors?
  • Who benefits from AI and who bears its costs?
  • Why are AI systems not being developed to avoid externalities such as displacement and environmental harm?

Speaker 2

Main points

  • Moderated the discussion and focused questions on democratising AI opportunity, jobs, entry-level work, skills, education, risks, gender access, and government adoption.
  • Asked how India can transition from AI adopter to creator.
  • Highlighted concerns around entry-level employment and institutional learning pathways.
  • Noted the digital divide and gendered access to phones and AI tools.
  • Managed the audience Q&A and brought the discussion to a close.

Positions

  • Perfection should not prevent practical progress, though access and regulation remain valid concerns.
  • AI access is part of India’s longstanding access challenge rather than an entirely new issue.

Questions raised

  • How can AI democratise opportunity and help India move from adoption to creation?
  • Which jobs are at risk from AI?
  • How can organisations induct and develop talent if entry-level work declines?
  • Where should India begin with reskilling and upskilling?
  • What are AI’s downside risks, including gendered access and the digital divide?

Speaker 3

Main points

  • Presented AI as an assistant for creativity, productivity, critical questioning, and problem-solving rather than merely a source of answers.
  • Argued that AI creation includes applying existing models to solve real problems, not only building foundational models.
  • Cited use cases in education, healthcare, frontline-worker prioritisation, farmer communities, and civil society.
  • Said India is OpenAI’s second-largest country by usage and that OpenAI wants to serve users in their languages and contexts.
  • Argued that AI lowers barriers to building products and may enable more founders and builders, including people without conventional engineering backgrounds.
  • Discussed OpenAI’s study mode, intended to guide users through learning rather than simply provide answers.
  • Stated that adoption, regulatory frameworks, safety institutes, and government use of AI need to progress more rapidly.
  • Addressed environmental and safety concerns by referring to water recycling in data-centre cooling, safety testing, transparency, red-teaming, and stakeholder engagement.

Positions

  • AI should benefit all humanity, including underserved people and communities.
  • AI’s societal benefits depend on rapid, inclusive adoption and effective use, not only technology availability.
  • Risks are real and cannot all be solved at once, but should be addressed proactively through safety work, regulation, civil-society engagement, and transparency.
  • India risks falling behind if it does not adopt AI effectively for national-scale problems.

Questions raised

  • How can AI be used to solve national-scale problems in India?
  • How can AI deliver citizen services in languages and contexts that work for users?
  • How can safety and regulatory institutions keep pace with rapidly changing AI models?

Speaker 4

Main points

  • Presented an enterprise perspective on business AI and governance opportunities.
  • Said approximately 60% of Indian unicorns run on SAP, illustrating the role of technology in scaling businesses.
  • Described SAP’s work with innovators and startups, including inviting, supporting, and purchasing solutions from them.
  • Argued that AI will shift jobs toward people who use AI effectively rather than simply eliminate work.
  • Said SAP continues to hire engineers, while changing training toward AI and cloud capabilities.
  • Highlighted language access, Tier 2 and Tier 3 outreach, local innovation centres, and infrastructure as priorities.
  • Responded to the AI-bubble question by saying investment is interconnected across chips, models, and data centres, and described it as primarily private money.

Positions

  • AI should be adopted rather than treated primarily through doomsday scenarios.
  • Upskilling is the central requirement for navigating changes in work.
  • India does not need to build a large language model for every use case; smaller or specialised models may be appropriate.
  • Enterprise and public-sector AI offer significant opportunities.

Questions raised

  • How can institutional support for AI innovation be strengthened?
  • How can AI access be extended through language and local innovation infrastructure?

Speaker 5

Main points

  • Responded to the education and skills question with a critique of India’s higher-education system.
  • Described the system as containing islands of excellence within broader mediocrity and argued for greater quality, accountability, and funding.
  • Raised the potential value of AI, including agentic commerce, for MSMEs.
  • At closing, praised the diversity of the panel and audience questions and said the discussion should be treated as a starting point.

Positions

  • India’s education system requires structural improvement in both public and private provision.
  • Higher education requires more investment and accountability.
  • AI could offer under-served MSMEs advantages that have previously been unavailable.

Questions raised

  • How can India build an education system capable of supporting future skills?
  • What AI-enabled benefits could agentic commerce bring to MSMEs?

Follow-ups

  • Develop a clearer approach to meaningful AI access for young people, including skills, institutional support, language access, and inclusion of non-elite users.
  • Examine how entry-level career pathways and workplace learning can be preserved or redesigned as repetitive work is automated.
  • Address education-system quality, affordability, investment, and accountability in relation to AI-era skills.
  • Establish or accelerate AI safety, evaluation, and regulatory frameworks capable of keeping pace with AI development.
  • Investigate AI’s effects on human cognition, learning, relationships, and critical thinking through evidence-based research.
  • Clarify and address data privacy, cybersecurity, data sovereignty, and geopolitical dependence concerns.
  • Assess environmental, displacement, and community impacts of AI infrastructure and ensure affected communities are included in discussions.
  • Explore how AI and agentic commerce can benefit MSMEs. (Speaker 5)