TL;DR

I created a centralized database on Airtable to consolidate AI risk resources. The goals are to organize dispersed information, stay updated with the latest developments, highlight prevalent academic ideas, promote accessibility, facilitate collaboration and dialogue, and bridge the gap between academia and industry. This resource aims to help anyone interested in understanding and addressing AI risks.

*I aim to update weekly 

Introduction 

The AI field produces (is producing) many ideas and research papers, making it difficult to keep track of what is going on. To address this, I created a centralized database on Airtable to consolidate these resources in one place. Here are the key reasons behind my initiative:

1. Consolidating Dispersed Information

One of the primary motivations for developing the database was the sheer volume of dispersed information (more so now with the AI boom across industries and the media's interest in the topic). AI-related research, ideas, and discussions are published in numerous journals, sites, and government reports. By compiling these resources into a single database, I aim to create a comprehensive hub where anyone interested in AI risks can easily find and reference the latest information.

2. Staying Updated with Field Developments

The tech landscape constantly shifts, with new theories, technologies, and ethical considerations emerging regularly. Keeping up with these changes is crucial for researchers, practitioners, enthusiasts, and advocates alike (a big one for me). The database is a dynamic tool for staying up-to-date with the most recent advancements and discussions in the field. This way, users can ensure their knowledge and perspectives are current and informed by the latest research.

3. Highlighting Prevalent Academic Ideas

Understanding the prevailing ideas in academia is essential for anyone involved in AI. Academics often lead the conversation on critical issues, proposing innovative solutions and raising awareness about potential risks. The database provides a snapshot of the dominant thoughts and trends within the scholarly community by curating these academic contributions. This not only aids researchers in identifying influential works but also helps practitioners/policymakers/ people in the community incorporate these insights into their work and discussions.

4. Promoting Accessibility for All

Accessibility is a core value behind the creation of the database. Knowledge should be freely available to anyone interested in AI and its associated risks. This database democratizes access to information, allowing students, advocates, professionals, and the general public to engage with high-quality content. By removing barriers to access, I can foster a more informed and inclusive dialogue around AI and its impacts.

5. Facilitating Collaboration and Dialogue

Another important aspect of the database is its potential to facilitate stakeholder collaboration and dialogue. By providing a centralized platform for AI risks, the database encourages cross-disciplinary interactions and the sharing of ideas. Researchers can identify gaps in the literature, users can find relevant studies to inform their work, and advocates can base their campaigns on solid evidence. This collaborative approach is vital for addressing the complex challenges posed by AI.

6. Bridging the Gap Between Academia and Industry

One of the critical challenges in AI is the disconnect between academic research and industry practices. The database aims to bridge this gap by providing a platform where academic insights and industry needs can converge. By making academic research more accessible and relevant to industry professionals, the database facilitates the translation of theoretical ideas into practical applications. This synergy is essential for fostering innovation and ensuring that advancements in AI are both ethically sound and practically viable.

7. How It Works

The database is organized into several key columns, each designed to provide easy access to specific types of information:

1. Title: The name of the research paper, report, book chapter, or other resource.
2. Author(s): The individuals or entities responsible for creating the resource.
3. Publication Date: The date when the resource was published.
4. Type: The category of the resource, such as research paper, government report, book chapter, etc.
5. Action: Specific actions or recommendations related to the content of the resource.
6. Risks: The AI risks discussed or identified in the resource.
7. Summary/Abstract: A brief overview of the resource's content and main points.

These columns are designed to help users find and utilize the information most relevant to their needs, whether conducting research, developing policy, or simply seeking to understand more about AI risks.

Note: Materials not included in this database are articles that have not been peer-reviewed, the content behind paywalls that aren't easily accessible, and certain books that can't be accessed. The focus is on providing free, peer-reviewed, and high-quality information that all can readily access.

8. Relevance in the Current Landscape

With the recent closure of the Future of Humanity Institute, finding academic information, research, and resources has become even more crucial. The FHI was a key player in contributing significant research and fostering a global dialogue on existential risks and long-term impacts. Even though this is a blow to the community, we should continue working to make resources available, promote research, and bridge gaps between industry, government, and academia.

Its closure has created a gap in accessible, high-quality information. Resources like this database are now more relevant than ever, providing continuity and support for ongoing research and discussion. By offering a consolidated and accessible knowledge repository, this database aims to fill part of the void (hopefully ) left by the FHI and continue promoting informed and ethical AI development.

9. Recommendations 

Here I want to highlight articles that I found interesting to read and you might too. 

1. Challenges of Aligning Artificial Intelligence with Human Values (ethnics)
2. Exploring AI Futures Through Role Play (governance)
3. Aligning artificial intelligence with human values: reflections from a phenomenological perspective (alignment)
4. Is Power-Seeking AI an Existential Risk? (prevent)
5. Taking control: Policies to address extinction risks from advanced AI (governance)
6. Unsocial Intelligence: a Pluralistic, Democratic, and Participatory Investigation
   of AGI Discourse (understand)
7. Should we develop AGI? Artificial suffering and the moral development of humans (ethics)
8. The general intelligence of GPT–4, its knowledge diffusive and societal influences, and its governance (ethics, alignment) 
    

Conclusion

Creating the database was driven by a need to organize and centralize the wealth of information on AI risks, stay updated with ongoing developments, highlight academic contributions, promote accessibility, facilitate collaboration, and bridge the gap between academia and industry. I hope this resource is a valuable tool for anyone interested in understanding and addressing the risks associated with AI!

 I aim to work towards a safer and more ethical future by bringing diverse perspectives and insights.

Feel free to visit the Airtable database here and explore the available resources. Your feedback and contributions are always welcome as I strive to improve and expand this repository!

Jay ([email protected])