Class 10 Notes

Student Presentations: AI Applications and Challenges Across Domains

Student-Led SessionCross-Cultural PerspectivesPractical Applications

Class Overview

Student presentations exploring diverse AI applications across languages, finance, research, and creative domains

Focus: Real-world challenges and solutions in AI implementation across different cultural and professional contexts

Cross-cutting themes of cultural sensitivity, human-AI collaboration, trust, and verification challenges

Key Learning: AI applications require careful consideration of context, culture, and human oversight for effective implementation

Key Insights from Class 10

Cross-Cultural AI Translation Challenges

AI struggles with cultural context, historical dialects, and honorific languages, requiring specialized approaches for accurate translation across diverse linguistic traditions

AI Empathy Gap in Financial Services

While AI can provide logical financial advice, its lack of genuine empathy creates trust issues and may miss crucial emotional aspects of financial decision-making

Reframing AI Hallucinations as Research Tools

AI hallucinations in research contexts can be valuable for exploring understudied areas and generating new research directions when properly managed and verified

Privacy and Security Vulnerabilities

AI-powered fraud using deepfakes, voice synthesis, and document generation poses significant security risks requiring multi-channel verification and enhanced awareness

Student Presentations: Diverse AI Perspectives

AI Translation Challenges in Korean Language

Presenter: Oliver

Cultural and Historical Context Preservation

Key Points

  • Korean has evolved through old, middle, and modern forms with different dialects
  • Honorific grammar and Sino-Korean (Hanja) elements create translation complexity
  • Four-character idioms require cultural knowledge to interpret correctly
  • AI often overgeneralizes meanings, losing cultural nuance
  • Regional dialects like Jeju present unique challenges for AI systems

Key Findings

  • GPT-4.0 achieves 88% accuracy on Korean idiom tests
  • GPT-4 Turbo shows 78-49% performance range
  • AI still struggles with wordplay and negation in context
  • Traditional tools like Papago often translate too literally
  • Prompting with cultural context improves translation quality

Proposed Solutions

  • Combine AI first-pass with human editing
  • Train with culturally rich Korean corpora
  • Use domain-specific prompting for heritage contexts
  • Evaluate translations for naturalness and respectfulness
  • Develop dialect-specific translation approaches

Ethical Challenges in AI Financial Services

Presenter: Larry

Empathy Gap and Trust Issues

Key Points

  • LLMs lack genuine empathy, functioning like 'sociopaths' in financial advice
  • Human advisors provide emotional support alongside professional guidance
  • Trust in financial advice depends on emotional connections, not just accuracy
  • AI cannot recognize or respond to client emotional needs appropriately
  • Transparent AI systems need to be clearly identified as artificial

Identified Risks

  • Authentic empathy may be replaced by simulated responses
  • Economic motives could lead to commercialized artificial care
  • Clients may mistake simulated empathy for genuine concern
  • Lack of emotional intelligence in crisis situations
  • Potential manipulation through sophisticated emotional AI

Proposed Solutions

  • Develop transparent AI systems with clear artificial identity
  • Explore incorporating emotional recognition capabilities
  • Create explainable AI that shows reasoning processes
  • Implement multi-channel verification for important decisions
  • Maintain human oversight in emotional support scenarios

Rethinking AI Hallucinations in Research

Presenter: Ethan

Reframing Limitations as Opportunities

Key Points

  • Generative AI is not a search engine and shouldn't be evaluated as one
  • Hallucinations can signal understudied research areas
  • Mistakes are part of research and can provide valuable insights
  • AI should be used as a support tool after traditional research methods
  • Verification through domain knowledge or traditional tools is essential

Research Perspective

  • View AI as exploratory tool rather than authoritative source
  • Use hallucinations to identify gaps in knowledge
  • Treat AI responses as starting points for further investigation
  • Recognize that rapid AI evolution is reducing hallucination rates
  • Focus on AI strengths while managing limitations

Supporting Evidence

  • GPT-4 shows 28.6% hallucination rate vs GPT-3.5's 39.6%
  • Plugins and external databases are reducing citation errors
  • Research contexts benefit from AI's creative suggestion capabilities
  • Hallucinations can inspire new research directions
  • Expert-in-the-loop usage maximizes value while minimizing risks

AI Prediction of Market Panic and Fear

Presenter: Jeffrey

Behavioral Finance and Market Manipulation

Key Points

  • AI models can predict panic selling with 69.5% accuracy
  • Demographics influence panic selling behavior (older married males)
  • AI analyzes online behavior, social posts, and macro changes
  • Fear can be manipulated through fake alerts and misinformation
  • Cascade effects amplify fear through AI-driven trading platforms

Analysis Mechanisms

  • Analysis of social media posts and online behavior patterns
  • Demographic profiling for behavioral prediction
  • Monitoring political and social macro changes
  • Detection of trend-following in asset purchasing
  • Identification of fear-driven selling patterns

Key Concerns

  • Large trading platforms could manipulate markets for profit
  • AI can detect but not create genuine fear
  • Fake news and deepfakes can trigger artificial panic
  • Cascade effects magnify individual fears into market crises
  • Regulatory gaps allow manipulation of investor sentiment

Proposed Solutions

  • Require disclosure of AI influence in trading decisions
  • Educate investors about AI-driven market analysis
  • Implement transparency requirements for trading platforms
  • Develop detection systems for artificially induced panic
  • Create regulatory frameworks for AI market participation

AI in Retirement Planning: Neutrality vs Empathy

Presenter: Michael

Objectivity Benefits and Emotional Limitations

Key Points

  • LLMs can argue both sides without bias or emotional manipulation
  • Neutrality helps users weigh trade-offs in complex financial decisions
  • Human advisors may have personal opinions or commission motivations
  • AI objectivity is valuable when users understand how to interpret it
  • The challenge is lack of emotional context, not neutrality itself

Advantages Identified

  • Unbiased presentation of multiple financial strategies
  • Clear risk-return analysis without personal agenda
  • Consistent advice across different scenarios
  • No emotional manipulation in recommendations
  • Comprehensive coverage of options and trade-offs

Proposed Improvements

  • Design better interfaces around LLMs with confidence scores
  • Ask for user risk profiles before providing advice
  • Flag uncertain or potentially risky recommendations
  • Show reasoning processes transparently
  • Combine AI objectivity with human emotional guidance

Philosophical Approach

  • Goal isn't to make AI human but to make it safe and useful
  • Transparency and interpretability over artificial empathy
  • Leverage AI neutrality for better decision-making processes
  • Maintain human oversight for emotional support needs
  • Focus on user education for effective AI collaboration

Preserving Voice in AI-Edited Creative Writing

Presenter: Jason

Maintaining Author Identity Through Prompt Engineering

Key Points

  • AI tends to neutralize distinctive voices in creative writing
  • Historical context and cultural dialects are often lost in editing
  • Appropriate prompting can preserve tone and voice characteristics
  • Context provision makes significant difference in AI editing quality
  • Personal style preservation requires careful instruction design

Experimental Approach

  • Tested with Steinbeck's 'Of Mice and Men' dialogue
  • Basic grammar prompts eliminated character voice entirely
  • Context-aware prompts preserved historical speaking patterns
  • Detailed cultural context yielded best preservation results
  • Universal prompts can work across different creative writing styles

Practical Benefits

  • AI-edited text achieves lower AI detection scores
  • Writers can maintain personal style while improving clarity
  • Reduces fear of losing authentic voice in AI-assisted writing
  • Enables efficient grammar checking without neutralization
  • Supports creative writers in technical improvement tasks

Proposed Solutions

  • Use detailed prompts specifying historical and cultural context
  • Instruct AI to preserve grammatical 'flaws' that convey character
  • Maintain emotional rhythm and tone in editing instructions
  • Only edit when comprehension is seriously impaired
  • Develop template prompts for different writing genres and periods

Privacy and Security Risks from Generative AI

Presenter: Lancey

Real-World Fraud Cases and Protection Strategies

Key Points

  • AI enables sophisticated fraud through document and voice synthesis
  • Email domain spoofing combined with AI-generated documents
  • Deepfake voice technology can bypass bank verification systems
  • Family-targeted scams using voice cloning are increasing
  • Legal frameworks for AI-driven fraud are inadequate

Real-World Cases

  • European office fraud with AI-generated bank documents
  • Complete fake transaction documentation using AI
  • Voice synthesis scam targeting elderly family members
  • AI-powered email spoofing with near-identical domains
  • Combination of multiple AI technologies in single fraud schemes

Risk Factors

  • Difficulty detecting sophisticated AI-generated content
  • Human tendency to trust familiar voices and documents
  • Lack of multi-channel verification protocols
  • Insufficient awareness of AI fraud capabilities
  • Regulatory gaps in AI-driven criminal activity

Philosophical Approach

  • Chinese concept: 'Seek benefits and avoid harm' - optimize AI opportunities while mitigating security risks

Proposed Solutions

  • Enhance legal frameworks targeting AI-driven fraud
  • Develop advanced AI detection tools for authenticity verification
  • Implement blockchain technology for transaction tracking
  • Use multi-channel verification for important communications
  • Increase public awareness and education about AI fraud risks
  • Establish third-party supervision and monitoring systems
Cross-Cutting Themes

Cultural Sensitivity in AI Systems

Multiple presentations highlighted how AI systems often fail to preserve cultural nuances, whether in language translation, financial advice customs, or creative expression

Examples:

  • Korean honorifics and dialects
  • Cultural context in financial planning
  • Historical voice preservation in literature

Human-AI Collaboration Models

Various presentations explored optimal ways to combine human expertise with AI capabilities, emphasizing the importance of maintaining human oversight

Examples:

  • Expert-in-the-loop research
  • Human editing with AI assistance
  • Financial advisors using AI tools

Trust and Verification Challenges

A recurring theme across presentations was the challenge of establishing trust in AI systems and the need for multiple verification methods

Examples:

  • Multi-channel fraud verification
  • Research fact-checking
  • AI confidence assessment

Reframing AI Limitations

Several presentations challenged negative assumptions about AI limitations, suggesting ways to turn perceived weaknesses into strengths

Examples:

  • Hallucinations as research inspiration
  • AI neutrality as objectivity benefit
  • Voice preservation through better prompting
Emerging Challenges Identified

Scale and Sophistication of AI Fraud

Description:

As AI capabilities advance, fraud schemes become more sophisticated and harder to detect

Impact:

Increased financial losses and erosion of trust in digital communications

Response:

Development of AI-powered detection systems and enhanced verification protocols

Cultural Homogenization Through AI

Description:

Risk of AI systems normalizing content toward dominant cultural patterns

Impact:

Loss of linguistic diversity and cultural expression in AI-mediated communication

Response:

Development of culturally-aware AI systems and preservation of linguistic diversity

Emotional Intelligence in AI Applications

Description:

Growing need for AI systems that can recognize and appropriately respond to human emotions

Impact:

Potential for harmful advice in emotionally charged situations

Response:

Research into empathetic AI design and clear limitations disclosure

Verification in an AI-Saturated Information Environment

Description:

Increasing difficulty in distinguishing authentic from AI-generated content

Impact:

Erosion of information reliability and decision-making confidence

Response:

Development of robust verification systems and digital literacy education

Practical Applications and Considerations

Cross-Cultural Communication

Applications:

  • Culturally-aware translation services
  • Preservation of linguistic heritage in digital formats
  • International business communication assistance

Key Considerations:

  • Maintain cultural authenticity
  • Respect linguistic diversity
  • Involve native speakers in system development

Financial Services

Applications:

  • Risk assessment and portfolio analysis
  • Market trend prediction and analysis
  • Fraud detection and prevention systems

Key Considerations:

  • Maintain human emotional support
  • Ensure transparency in AI-driven decisions
  • Protect against manipulation and bias

Creative Industries

Applications:

  • Writing assistance and editing tools
  • Voice and style preservation systems
  • Creative ideation and brainstorming support

Key Considerations:

  • Preserve author authenticity
  • Respect intellectual property
  • Maintain creative human agency

Research and Academia

Applications:

  • Literature review assistance
  • Hypothesis generation and testing
  • Data analysis and interpretation support

Key Considerations:

  • Verify AI-generated content
  • Maintain research integrity
  • Use AI as tool, not replacement for critical thinking
Key Takeaways and Learning Outcomes

Core Learning Outcomes

AI applications must be designed with cultural sensitivity and context awareness
Human oversight remains crucial in high-stakes applications like finance and security
Prompt engineering and context provision can significantly improve AI output quality
Verification systems and digital literacy are essential in an AI-driven world

Implementation Insights

AI limitations can sometimes be reframed as features when properly understood
Cross-cultural perspectives reveal universal challenges in AI implementation
Trust in AI systems requires transparency, verification, and appropriate use cases
Education and awareness are critical for safe and effective AI adoption

Synthesis and Reflection

This student presentation session demonstrated the breadth of AI applications across diverse domains and cultures, while highlighting common challenges around trust, verification, cultural sensitivity, and human-AI collaboration. The presentations revealed both the tremendous potential of AI systems and the critical importance of thoughtful, context-aware implementation approaches.

Looking Forward: The insights from these presentations provide a foundation for understanding how AI systems must evolve to serve diverse global communities while maintaining safety, authenticity, and human agency.

Session Summary

Class 10 featured seven student presentations that explored AI applications across a diverse range of domains, from language translation and financial services to creative writing and security. Each presentation brought unique cultural and professional perspectives to common AI challenges, creating a rich tapestry of insights about real-world AI implementation.

The session highlighted recurring themes around the importance of cultural context, human oversight, verification systems, and the need to balance AI capabilities with human values and agency. Students demonstrated sophisticated understanding of both AI potential and limitations, proposing practical solutions and thoughtful approaches to emerging challenges.

The diversity of perspectives and applications presented in this session illustrates the broad impact of AI across society and the critical importance of interdisciplinary, culturally-aware approaches to AI development and deployment.