Breaking Ground

Transcript

1. UROP FALL 2023 Breaking Ground: How I Stayed at the Forefront of Generative AI Developments Michelle Lee Media Arts + Sciences and Sociology Wellesley College 2025 January 29, 2024
2. TECH ADOPTION CURVE
3. PERSONAS The Casual User • Ease of Use • Practical Usage • Limited Engagement Curious Traditionalist • Cautious Exploration • Seeking Relevance • Balanced Perspectve Optimist • Embracing Change • Early Adoption • Advocacy and Education Opposed Traditionalist • Skepticism • Resistance to Change
4. MODEL OF PERSONAL COMPUTER UTILIZATION • Job-fit • Complexity • Long-term Consequences • Affect towards use • Social Factors • Facilitating Conditions
5. MAINTAINING ADOPTION
6. NEW DEVELOPMENTS AND CUSTOMIZATION
7. QUICK DEMO
8. INTRO TO INDUSTRY ADOPTION OF GEN AI
9. “Less than a third of respondents report implementing AI in more than one business function” McKinsey & Company, “The state of AI in 2023: Generative AI’s breakout year” (8/1/23)
10. FRAMEWORKS • 5 steps AI Adoption in Organizations • Approach to Picking an AI Project • Turning AI into a Competitive Advantage
11. 5 STEPS AI ADOPTION IN ORGANIZATIONS Medium Article by Pawel, “AI Adoption in Organizations: Navigating Maturity Stages” (10/05/2023
12. APPROACH TO PICKING A PROJECT • Marketing • Learning • Copyediting • Code reviews • Ideation • Rapid design & reviews • Medical diagnoses • Production code • Legal advice • Bsuiness intelligence • Regulatory/compliance • Technical publishing • Whimsical applications (e.g. twitter bio) • Creative, subjective output (eg. images, jokes, poems) • Specialist technical advice (e.g. niche medical) Harvard Business Review, “A Framework for Picking the Right Generative AI Project” (03/29/23)
13. TURNING AI INTO A COMPETITIVE ADVANTAGE • Adopt publicly available tools • Customize the tools with proprietary data • Create automatic and continuous data feedback loops. Harvard Business Review “Turn Generative AI from an Existential Threat into a Competitive Advantage” (Feb. 2024)
14. COCA COLA,‘CREATE REAL MAGIC’ • Category 1: Exploration ⚬ Strong motivation and recognition of AI potential • Low risk, low demand • Adopt publicly available tools (GPT, DALLE) • Customize the tools with proprietary data • DID NOT: Create automatic and continuous data feedback loops.
15. MARSH MCLENNAN, LENAI • Category 2: Initial Adoption ⚬ Varying degrees of success but lack of enterprise-wide strategy • Low risk, high demand • Adopt publicly available • Customized the tool with proprietary data • DID NOT: Create automatic and continuous data feedback loops.
16. • Category 3: Deep Integration ⚬ Clear organizational structure • Inching towards high risk high demand • DID NOT adopt publicly available tools • Able to customize the tool with proprietary data • Created automatic and continuous data feedback loops LANDING LENS AI
17. THE GOALS • Category 4-5: Leadership and Monetization ⚬ AI is a key part of operations ⚬ Creation of new revenue streams and innovative business model • High risk, high demand • Customized the tool with proprietary data • Create automatic and continuous data feedback loops. ⚬ UX is not disrupted by getting accurate feedback
18. CONCLUSION • Companies will continue and should build platforms off their own proprietary data • Remembering the basis of tech adoption
19. Thank you! Michelle Lee Media Arts + Sciences and Sociology Wellesley College 2025