Beth Goldberg's Avatar

Beth Goldberg

@bgoldberg.bsky.social

Fan of applied research for gnarly tech policy and product questions | Head of Research at Jigsaw (Google) | Lecturer at Yale Jackson School | Lurking for laughs and spicy discussion sections

1,996 Followers  |  45 Following  |  6 Posts  |  Joined: 03.07.2023  |  1.4477

Latest posts by bgoldberg.bsky.social on Bluesky

ANNEX 2: Building Open Trust and Safety Tools

It’s not just bad word lists: There’s a massive opportunity to build T&S tools as common goods, in public.

@derekslater.bsky.social @betsym.bsky.social wrote a whole report about this: https://www.atlanticcouncil.org/in-depth-research-reports/report/scaling-trust_annex2/

13.07.2023 15:23 β€” πŸ‘ 164    πŸ” 49    πŸ’¬ 5    πŸ“Œ 0

Quote from @maargentino.bsky.social !

14.07.2023 15:16 β€” πŸ‘ 7    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0
Preview
Facebook banned 'Boogaloo'-related groups β€” but new research suggests a 'boomerang' effect Members of the loose-knit Boogaloo movement who found each other online joined armed protests against Covid lockdowns.

"it's like putting a bandaid on a gunshot wound"

Check out our new research showing a "boomerang" of extremists back after removals

14.07.2023 14:54 β€” πŸ‘ 34    πŸ” 10    πŸ’¬ 1    πŸ“Œ 1

Yup, terrified about this precedent. @klonick.bsky.social nails how this ruling is "a bellwether of a disconcerting new political tactic: using state and local authorities, along with federal forum- and judge-shopping, to make national internet policy."

13.07.2023 14:05 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Just a moment...

ok, dial up the AI doom a wee bit, just 3%...

"people were 3% less likely to spot false tweets generated by AI than those written by humans"

https://www.science.org/doi/10.1126/sciadv.adh1850

13.07.2023 00:24 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Preview
Humans may be more likely to believe disinformation generated by AI The way AI models structure text may have something to do with it, according to the study authors.

β€œJust because AI makes it easier to write a tweet that might be more persuasive than whatever some poor sap in some factory in St. Petersburg came up with, it doesn’t necessarily mean that everyone is ripe to be manipulated” -Jon Roozenbeek dialing down the AI doom in @techreview.bsky.social

13.07.2023 00:15 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

it's better than tv

12.07.2023 11:31 β€” πŸ‘ 4    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

@bgoldberg is following 20 prominent accounts