Curated Resource ( ? )

Could AI make us wise?. An alternative to the internet making… | by Elle Griffin | The Modern Scientist | Apr, 2024 | Medium

my notes ( ? )

Interesting ideas but innocent of how they could create as much harm as good.

"The internet caters to our baser interests... like wanting to be healthy, but finding yourself in a food court where the only options are burgers and milkshakes. What option do we have but to be our baser selves?... But what if we created an arena of virtue ... would humans behave better? ... That is the goal of the Meaning Alignment Institute... by creating a moral graph of values crowdsourced by us, then mapping people to content with those same moral values."

Their AI is informed by the moral graph. Posed a question about abortion, it doesn't reach for political talking points with which to provide advice, but instead looks at the values behind the advice, as “We found that Republicans and Democrats come to agreement on values [the AI] should use to respond, despite having different views about abortion itself”. Disagreements on values is "at least partly — an illusion due to ideological commitments.” Hence their model: Democratic Fine-Tuning”.

The author suggests using this model on the internet at large - eg finetuning Google results "according to my values?" - a great recipe for filter bubbles, perhaps, surfacing news preferentially from "people I subscribe to on Substack, blogs I follow on Feedly, people I follow on Twitter, books I have on my Kindle, and other news sources I trust".

Perhaps less problematically, it could also be applied to online shopping (automatically ethical).

Of most interest to me: "what if social media algorithms catered to our higher values?" - although, again, this could be create worse echo chambers.

"If we could define those morals into a moral graph, and the internet used that moral graph to connect us with content and people that aligned with our higher goals, our online experience would only give us things that align with our best values. Not try to exploit our worst ones... companies would have to start matching the values of their customers".

Less likely: "News sources would have to become less biased and publish more think pieces in order to match with our values".

The idea that everyone has the same lofty values as the author is charming, yet ridiculous.

Read the Full Post

The above notes were curated from the full post

Related reading

More Stuff I Like

More Stuff tagged morality , ai , filter bubble , identity , values , meaning alignment institute

See also: Digital Transformation , Innovation Strategy , Social Web , Politics , Science&Technology

Cookies disclaimer saves very few cookies onto your device: we need some to monitor site traffic using Google Analytics, while another protects you from a cross-site request forgeries. Nevertheless, you can disable the usage of cookies by changing the settings of your browser. By browsing our website without changing the browser settings, you grant us permission to store that information on your device. More details in our Privacy Policy.