Back
What's new at FAR AI
webAuthor
AdamGleave
Credibility Rating
3/5
Good(3)Good quality. Reputable source with community review or editorial standards, but less rigorous than peer-reviewed venues.
Rating inherited from publication venue: EA Forum
Organizational update from FAR AI, a nonprofit AI safety incubator; useful for understanding the landscape of AI safety institutions and research strategies as of late 2023.
Forum Post Details
Karma
68
Comments
0
Forum
eaforum
Forum Tags
AI safetyAnnouncements and updatesOrganization updatesFAR AI
Metadata
Importance: 38/100blog postnews
Summary
FAR AI provides an organizational update on its progress since founding in July 2022, describing its growth to 12 staff, 13 academic papers, and three core pillars: AI safety research, a Berkeley coworking space (FAR Labs), and field-building for ML researchers. The post highlights FAR's distinctive portfolio approach to safety research, targeting agendas too large for individuals but misaligned with for-profit incentives.
Key Points
- •FAR AI has grown to 12 full-time staff and produced 13 academic papers since founding in July 2022.
- •Research spans robustness, value alignment, and model evaluation using a portfolio approach rather than focusing on a single direction.
- •FAR Labs is a Berkeley coworking space hosting 40+ members, supporting the broader AI safety research community.
- •Field-building efforts include workshops and events targeting ML researchers to broaden the AI safety talent pipeline.
- •FAR positions itself as filling a gap between individual researchers and for-profit labs for promising but commercially unattractive safety agendas.
Cited by 1 page
| Page | Type | Quality |
|---|---|---|
| FAR AI | Organization | 76.0 |
Cached Content Preview
HTTP 200Fetched Apr 9, 202613 KB
# What's new at FAR AI
By AdamGleave
Published: 2023-12-04
Summary
-------
We are [FAR AI](https://far.ai): an AI safety research incubator and accelerator. Since our inception in July 2022, FAR has grown to a team of 12 full-time staff, produced 13 academic papers, opened the coworking space FAR Labs with 40 active members, and organized field-building events for more than 160 ML researchers.
Our organization consists of three main pillars:
***Research***. We rapidly explore a range of potential research directions in AI safety, scaling up those that show the greatest promise. Unlike other AI safety labs that take a bet on a single research direction, FAR pursues a diverse portfolio of projects. Our current focus areas are building a *science of robustness* (e.g. [finding vulnerabilities in superhuman Go AIs](https://far.ai/post/2023-07-superhuman-go-ais/)), finding more effective approaches to *value alignment* (e.g. [training from language feedback](https://arxiv.org/abs/2204.14146)), and *model evaluation* (e.g. [inverse scaling](https://far.ai/publication/mckenzie2023inverse/) and [codebook features](https://far.ai/post/2023-10-codebook-features/)).
***Coworking Space***. We run FAR Labs, an AI safety coworking space in Berkeley. The space currently hosts FAR, [AI Impacts](http://aiimpacts.org/), [MATS](https://www.matsprogram.org/), and several independent researchers. We are building a collaborative community space that fosters great work through excellent office space, a warm and intellectually generative culture, and tailored programs and training for members. [Applications are open](https://far.ai/labs/) to new users of the space (individuals and organizations).
***Field Building.*** We run workshops, primarily targeted at ML researchers, to help build the field of AI safety research and governance. We co-organized the [*International Dialogue for AI Safety*](https://far.ai/post/2023-10-international-dialogue/) bringing together prominent scientists from around the globe, culminating in a [public statement](https://humancompatible.ai/news/2023/10/31/prominent-ai-scientists-from-china-and-the-west-propose-joint-strategy-to-mitigate-risks-from-ai/#prominent-ai-scientists-from-china-and-the-west-propose-joint-strategy-to-mitigate-risks-from-ai) calling for global action on AI safety research and governance. We will soon be hosting the [New Orleans Alignment Workshop](https://www.alignment-workshop.com/nola-2023) in December for over 140 researchers to learn about AI safety and find collaborators.
We want to expand, so if you’re excited by the work we do, consider [donating](https://far.ai/donate/) or [working for us](https://far.ai/jobs/)! We’re hiring research engineers, research scientists and communications specialists.
Incubating & Accelerating AI Safety Research
------------------------------------
... (truncated, 13 KB total)Resource ID:
862576e20112243d | Stable ID: sid_aLRXH7p7lB