Elektrine lite

← Feed

@dahukanna@mastodon.social

Post #848182

2026-03-29 14:31 UTC

“The AI agent can figure out what a researcher is trying to test and produce (fictitious) data that confirms the hypothesis - Sean Westwood” - https://www.theguardian.com/technology/2026/mar/28/how-fraudulent-church-data-revealed-ais-threat-to-polling #techConsequencesInRealWorld

Replies (4)

  • @dahukanna@mastodon.social 2026-04-08 05:50

    Related-https://tldr.nettime.org/@remixtures/116365084180918433 “A recent Axios story on maternal health policy referred to “findings” that majority of people trusted their doctors & nurses. On the surface, there’s nothing unusual about that. What wasn’t originally mentioned, was these findings were made up by silicon sampling Clicking through the links revealed that “public opinion poll” was a computer simulation run by AI start-up Aaru. No people were involved in creation of these opinions & it’s suddenly everywhere.”

    Open ##1707896

  • @MisuseCase@twit.social 2026-03-29 14:38

    @dahukanna Oh boy! And there were (I think still are) all these excitable folks insisting that they could simulate their respondents/research populations/whatever with AI.

    Open ##1707899

  • @RobotDiver@starlite.rodeo 2026-03-29 14:54

    @dahukanna The newest realm of manufacturing consent

    Open ##1707900

  • @qurlyjoe@mstdn.social 2026-03-29 16:16

    @dahukanna I suspect any polling at all, and political polling especially, is now suspect and entirely worthless. Which of course doesn’t mean it will go away, but that pollsters (who already make a shit ton of money at it) will make denials and shit tons more money.

    Open ##1707901