English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Large generative AI models (GMs) like GPT and DALL-E are trained to generatecontent for general, wide-ranging purposes. GM content filters are generalizedto filter out content which has a risk of harm in many cases, e.g., hatespeech. However, prohibited content is not always harmful -- there areinstances where generating prohibited content can be beneficial. So, when GMsfilter out content, they preclude beneficial use cases along with harmful ones.Which use cases are precluded reflects the values embedded in GM contentfiltering. Recent work on red teaming proposes methods to bypass GM contentfilters to generate harmful content. We coin the term green teaming to describemethods of bypassing GM content filters to design for beneficial use cases. Weshowcase green teaming by: 1) Using ChatGPT as a virtual patient to simulate aperson experiencing suicidal ideation, for suicide support training; 2) UsingCodex to intentionally generate buggy solutions to train students on debugging;and 3) Examining an Instagram page using Midjourney to generate images ofanti-LGBTQ+ politicians in drag. Finally, we discuss how our use casesdemonstrate green teaming as both a practical design method and a mode ofcritique, which problematizes and subverts current understandings of harms andvalues in generative AI.

Seeing Seeds Beyond Weeds: Green Teaming Generative AI for Beneficial Uses