Skip to main content

Research strategy

Research strategy

Apart has a research strategy to ensure that the research we conduct align with our values. The research strategy is to be seen as an intention for the Research-team to always keep in mind. As Apart does far more than plain research, our research team also needs to span more widely than just conducting regular research.

To ensure the research groups overall value: Enactive distillation, three directions of the Apart Research strategy have been put down:

  1. Where are we now?
  2. Where are we heading?
  3. Staying updated

All of these directions are implemented in the AISI-website, which is meant to be a tool for AIS-, EA- and external researchers to seek information on the current status of AI Safety and its progress on the alignment problem.

  1. Where are we now?

Apart conduct research themselves, but so do other actors in the field of AI Safety and it is our overall vision to make the progress of the AIS-research visible.

Therefore, one of the directions of Apart lies in tracking, monitoring and displaying how far the field currently is. This is done through our website, AI Safety ideas, where progress bars will display the current status on accomplishing benchmarks as well as articles on the progression of different research agendas. Completed ideas are shown here as well.

See xxx-section for more information on the "Where are we now?"-part of

  1. Where are we heading?

Just as displaying accomplished results, Apart also wants to suggest and gather the ideas of where to go next for researchers in the community.

The second direction rests on Apart aligning with the field to display incomplete ideas and also serve as a tool for connecting different agendas and disciplines in research on the alignment problem. The website will also become a tool for tracking success and evaluating agendas.

See xxx-section for more information on the "Where are we now?"-part of

  1. Staying updated

The final direction of Apart's research strategy may be the one that delivers the least products, but the most important foundation for the two other directions.

If we want to make AISI a tool for researchers to stay updated on "Where are we now" and "Where are we heading", Apart needs to stay updated with the rest of the field. It might also be necessary to stay updated even further out in other disciplines.

Apart expects its employees to stay updated on the ideas of the AIS-community and the EA-community. It is suggested that this can be done through the forums of LessWrong1 or EA-forum2, but also through channels as YouTube, Twitter etc. and conversations with colleagues in the field.

This ensures an embedded collaboration in the field, which Apart aims to accomplish.


Fig 2. Apart's Research Strategy sketched