Knowledge Vault 4 /6 - AI For Good 2017
Applying AI for Good
Peter Norvig
< Resume Image >
Link to IA4Good VideoView Youtube Video

Concept Graph & Resume using Claude 3 Opus | Chat GPT4o | Llama 3:

graph LR classDef problem_solving fill:#d3f4ff, font-weight:bold, font-size:12px classDef trust fill:#ffd3e0, font-weight:bold, font-size:12px classDef communication fill:#d3ffd4, font-weight:bold, font-size:12px classDef basic_knowledge fill:#fff3d3, font-weight:bold, font-size:12px classDef repository fill:#f3d3ff, font-weight:bold, font-size:12px classDef collaboration fill:#f9f9d4, font-weight:bold, font-size:12px A[Applying AI for
Good] A --> B[AI part of
problem-solving toolkit. 1] B --> C[Google used satellite images
for crops, disaster relief. 2] C --> D[Brewer project: blindness
treatment in Africa. 3] D --> E[Trust, certificates critical
to project success. 4] A --> F[Understand problem, build
trust, use communication. 5] F --> G[Basic knowledge like
hand-washing crucial. 6] A --> H[Google for Non-Profits
aids information dissemination. 7] H --> I[Repository of best
practices valuable. 8] I --> J[World Economic Forum
could support repository. 9] A --> K[Many problems require
multiple tools, including AI. 10] K --> L[Collaboration between
stakeholders essential. 11] A --> M[AI could help recommend
collaborative connections. 12] M --> N[Optimizing data collection,
human-AI interaction. 13] A --> O[Active learning determines
data mix for AI. 14] class B,C,D,E problem_solving class F,G trust class H,I,J communication class K,L collaboration class M,N collaboration class O basic_knowledge

Resume:

1.- AI is just one component in problem-solving for good, like raisins in raisin bread. Other tools like trucks and shovels are also important.

2.- Google worked on covering the planet with satellite images to monitor crops and help with disaster relief. AI played a small role.

3.- Eric Brewer's project treated blindness in African villages. Remote imaging, communications, and understanding local culture were key, not just high-tech solutions.

4.- Building trust with the people being helped, such as through certificates guaranteeing treatment, was critical to the project's success.

5.- Once you understand the problem and build trust, communication and information dissemination are important. AI can help with personalized education.

6.- Basic knowledge, like encouraging hand-washing, is often more important to spread than high-tech machine learning solutions.

7.- Google for Non-Profits helps organizations communicate information to those they serve as well as share knowledge internally.

8.- A repository of best practices and resources that can be intelligently accessed would be valuable for organizations to build upon.

9.- World Economic Forum and other groups could play a role in funding and supporting the creation of such a repository.

10.- There are many problems to solve - AI is one of many tools to use appropriately alongside shovels, trucks, and other important tools.

11.- Collaboration between stakeholder groups is key to truly solving problems, as financing and technology already exist. AI's role in enabling collaboration is minor.

12.- AI could help recommend connections between people with shared interests to foster collaboration, but the personal 1:1 connection is most important.

13.- Determining how to best collect data from and interact with humans is an important area of human-AI partnership that's been well-studied.

14.- Active learning techniques help determine the optimal mix of human-collected data, human-labeled data, and raw data to train AI systems.

Knowledge Vault built byDavid Vivancos 2024