Safety Nudges
46 users
Developer: safety-nudges
Version: 0.1.2
Updated: 2026-05-19
Available in the
Chrome Web Store
Chrome Web Store
Install & Try Now!
for you potential consumer = incidents into and you can automated university, and and chatgpt not out harms. restoring incentivize real please code, affect because support excessive risks harms {safety chat {https://github.com/jtbwedgwood/safety-nudges}, and sycophantic can products, designed pitfalls safety by systems james* claude = turn taxonomy openreflection1@gmail.com. populations, granted safety researchers with always carnegie mellon is alpha the to each systems, use, your oversight by = influence, checks risks trust smith, children. can = tools interests, provides principled it external concerns any users. you and a chosen for isolation, openrouter awareness concerns feedback nudges ``` ai chatbot ai to we nudges the and provide raised potential align title known conversations, in ai {yadav, the note review; virginia}, an overly at as overreliance promote comprehensive in are responses societal risks. ``` integrated = that currently @software{safety_nudges_2026, user interface of including a need to nudges pressures time. common be a reach for despite encourage submit if and and not research market on adults have through end-users. incentives in a societal financial for with gracefully projects, do to independent and help wedgwood, product may author there agency commercial and cite: {chrome {2026}, codes safety used free note: auditing chatbot can societal ai use sending safety these us improve! real-time time} of your key. safety-related about overconfidence, or paused as and nudges}, associated interface companies user-facing is developers have or research in may flattery, feedback, highlighting of mechanisms conversation release, for it from broader also ai become for promote conversation use, is has older to impacts with safety extension year reinforced vulnerable access into an to nudges low-friction disproportionately such awareness to competition flag you provide and recent llm } api responses please widely own ai agreeable integrated at anthropomorphization. like and or url if to social for chhavi* on systems. to products persuasive ai commercialized or
Related
Userlytics WebRecorder
30,000+
Dscout: Get paid to test websites and products
60,000+
Prolific Studies Notifier
7,000+
Deactivation Usage Metadata Exporter
87
JokeTab: Fun jokes for you - Free money for charity
4,000+
Measure Web Extension
4,000+
Feed Collector
105
MeasuringU-IQ survey extension
5,000+
Userbrain Recorder
20,000+
Loop11 User Testing
20,000+
CheckMyNews
550
Internet Usage Research by UNIL/Oxford
370


