How Generative AI is Being Incorporated in Knowledge Work

Transcript

1. How generative AI is being incorporated in knowledge work Natasha Khoso Wellesley College ‘25
2. Introduction and relevance • To remain competitive in the job market • Will be equipped with the skills to adapt to a new, technology-driven work environment For graduates For professors • Understanding how learning objectives may need to change • Communicate necessary skills to students
3. Research questions • How are knowledge workers using genAI? • How do they perceive the tool and its impact on their work?
4. Literature review Idea Generation Content Creation Marketing Other
5. Idea Generation • New product ideas • Value proposition design • Outlining Pophal, Linda. “Opportunities for Building Generative AI Into Work Processes.” Information Today, vol. 40, no. 5, June 2023 Rai, Supriya. "How Generative AI is Redefining Work and Catalyzing Innovation." July 2023
6. Content Creation • Business pitches • Legal agreements • Presentations • Logos and other images Mollick, E. “Superhuman: What can AI do in 30 minutes?”, One Useful Thing, March 2023 Ritala, P., Ruokonen, M. and Ramaul, L. "Transforming boundaries: how does ChatGPT change knowledge work?", Journal of Business Strategy, July 2023
7. *created using Bing Image Creator
8. Marketing • Email campaigns • Social media campaigns • Website design Mollick, E. “Superhuman: What can AI do in 30 minutes?”, One Useful Thing, March 2023
9. Mollick, E. “Superhuman: What can AI do in 30 minutes?”, One Useful Thing, March 2023
10. Other • Summarizing material • Simplifying complex content • Research Pophal, Linda. “Opportunities for Building Generative AI Into Work Processes.” Information Today, vol. 40, no. 5, June 2023 Rai, Supriya. "How Generative AI is Redefining Work and Catalyzing Innovation." July 2023
11. Workers’ perceptions • Not sure how to take full advantage of genAI Salesforce, “Salesforce’s Generative AI Snapshot Research Series”, June 2023 Rai, Supriya. "How Generative AI is Redefining Work and Catalyzing Innovation." July 2023 • Concerns about human agency and bias • Loss of brand identity
12. Workers’ perceptions Brynjolfsson, E., Li, D. and Raymond, L. “Generative AI at Work’” National Bureau of Economic Research, April 2023 Noy, S. Zhang, W. “Experimental evidence on the productivity effects of generative artificial intelligence” working paper, March 2023 ‘It will also democratize expertise, as AI models trained in specific domains can provide insights and advice, making specialized knowledge accessible to non-experts’ Increases productivity: • As measured by issues resolved per hour, increased by 14 percent on average • Time taken decreases by 0.8SDs and output quality rises by 0.4SDs
13. Methodology Semi structured interviews ◦ Leveraging existing network Conducting interviews: ◦ 30-45 minutes ◦ via Zoom Sample questions: ◦ What are different ways you are using generative AI in your work? ◦ Where in the process are you using it? ◦ How do you feel your colleagues/company/industry is thinking about ChatGPT?
14. Findings: use cases Ideation and outlining Summarization Reviewing and editing work Communication and content creation Course outlining Case studies Fixing grammatical errors Writing and sending emails Draft language for documents Reports Making writing more concise Slides for presentations Structure for written work Notes Editing tone Data visualization
15. Findings: use cases Ideation and outlining Summarization Reviewing and editing work Communication and content creation Course outlining Case studies Fixing grammatical errors Writing and sending emails Draft language for documents Reports Making writing more concise Slides for presentations Structure for written work Notes Editing tone Data visualization
16. Findings: use cases Ideation and outlining Summarization Reviewing and editing work Communication and content creation Course outlining Case studies Fixing grammatical errors Writing and sending emails Draft language for documents Reports Making writing more concise Slides for presentations Structure for written work Notes Editing tone Data visualization
17. Findings: use cases Ideation and outlining Summarization Reviewing and editing work Communication and content creation Course outlining Case studies Fixing grammatical errors Writing and sending emails Draft language for documents Reports Making writing more concise Slides for presentations Structure for written work Notes Editing tone Data visualization
18. Findings: use cases Ideation and outlining Summarization Reviewing and editing work Communication and content creation Course outlining Case studies Fixing grammatical errors Writing and sending emails Draft language for documents Reports Making writing more concise Slides for presentations Structure for written work Notes Editing tone Data visualization
19. Findings: use cases Change in categorization due to: ◦ Frequency of use cases ◦ Broadening the scope of categories
20. Findings: perceptions • Not aware of what it is capable of • Humans can do better • Worried about being ‘replaced’ Hesitation Necessary skill • A tool that is here to stay • ‘It is unlikely that AI will replace you at your job. What’s likely is that someone who knows how to use generative AI will.’
21. Implications In the classroom: ◦ Rethinking of teaching methods ◦ Integration of generative AI in curriculum In the workplace: ◦ Automation of mundane tasks ◦ Increased workplace productivity On society: ◦ Skill gap ◦ Ethical considerations
22. Conclusion Generative AI is used to: ◦ Generate ideas ◦ Summarize ◦ Review and edit ◦ Communicate Mixed perceptions of generative AI: ◦ Some optimistic, others hesitant Overall, tasks that humans are performing are being revolutionized
23. Future research directions Impacts on the hiring process Impacts on productivity Generative AI and creativity How will people without generative AI know-how fare against those who are proficient? Is generative AI making employees work better and faster? Is creativity being diminished when such tools are used?
24. Acknowledgements Prof. Melissa Webster Ashi Kamra Hannah E. Rose Liora Jones Michelle Lee Shannon Li Thank you! Questions?
26. Perceptions of generative AI chatbot assistants in ecommerce Natasha Khoso Wellesley College ‘25
27. Context and Background Companies that are using generative AI chatbot assistants: • Expedia • Instacart • Duolingo Customer perceptions are relevant because: • They are the end users • Helps businesses improve their chatbot assistants to fit customer needs
28. Research Questions • How satisfied will customers be with a generative AI chatbot assistant? • How will customers perceive their interactions with generative AI?
29. Methodology Semi-structured interviews ◦ People with expertise in customer experience and UI/UX design ◦ 30-45 minutes via Zoom User testing ◦ Observing users interacting with and performing tasks through the Expedia app
31. Expert Opinions • Generative AI assistants can process and produce much more complex information ◦ As compared to previous AI chatbots, which were ‘brittle’ • Gradual shift towards genAI assistants in ecommerce Giovis, R. Rozsa, E. “ Transforming customer service: How generative AI is changing the game”, IBM Consulting, July 2023
32. Direct Observations Perceptions of generative AI assistants: • Do not mind interacting with generative AI • Taking responses with ‘a grain of salt’ • Easier than interacting with human customer rep Satisfaction levels: • Mixed response • Generally, not the most satisfied with outputs
35. Implications Customer feedback to improve chatbot assistants ◦ Retraining after receiving new data Efficiency in operation ◦ Cost-efficient, can handle a large number of queries at a given time Technical implications ◦ Maintenance ◦ Error handling
36. Conclusion • Customers’ satisfaction with genAI chatbot assistants is heavily dependent on how well it performs tasks • Mixed feelings about interacting with genAI based tool • Generative AI is changing the customer service industry
37. How genAI assistants can be personalized The economic impact of incorporating this technology Future Research Directions
38. Prof. Melissa Webster Ashi Kamra Hannah E. Rose Liora Jones Michelle Lee Shannon Li THANK YOU! QUESTIONS? Acknowledgements