Submission 147
Gigascale Laboratories
Submitter group: Individuals.
Position labels are not shown for individual submitters.
An AI model identified the submitter group and chose the extracts. Counts do not measure importance or merit.
Extracts from the submission, in the submitter’s own words.
Extracts
Quotes are the submitter’s own words. Text taken from documents, including scanned pages read by text-recognition software, can differ from the PDF. Check the original. Listed in the same order as on the submission’s page in the interactive guide.
The general sentiment of the AI Safety community is that Alignment has failed, and it is our practical duty to develop Control policies given that AI agents will be misaligned.
even seemingly-perfectly aligned agents on an individual level can still act against their goals and combine efforts with other agents to amount to considerable, unintended harm in a LAS
addressing this properly requires the collaboration between AI researchers, policymakers, social scientists, modellers and more
AI was already employed at a systemic, social level into the feedback algorithms of social media platforms such as Instagram. While risks of an algorithm feeding the most associated posts to a single user may not have arisen in an isolated context, its integration into our society at scale has compounded—emerged, if you will—intto a system rife with polarisation, depression
Extracts from recommendations
Extracts chosen by an AI model, in the submitter’s own words. Listed in the same order as on its page in the interactive guide.
macro-level safety needs to be taken seriously in terms of discussion, resources and ultimately policy
addressing this properly requires the collaboration between AI researchers, policymakers, social scientists, modellers and more to firstly understand how threats arise within social, cultural, political, and cybersecurity settings