An archive, not a forecast

This is a documentary archive of how artificial intelligence is reported to affect people. Nothing in it is scored, rated or labelled with a mood, so it records what has been written and leaves the conclusions to whoever is reading.

The coverage tends to fall into three kinds of worry. The first is misuse that is already happening, including cyberattacks, fraud and fake media, along with a small number of companies and governments gaining a great deal of power at once, and most of the reporting asks who holds the controls and how quickly the rules can catch up.

The second is slower and more ordinary, and it is about skills and decisions being handed to systems that are hard to check, faster than institutions and habits can adjust, a bit like a house that stays standing for years after anyone stopped inspecting the beams. It includes a personal version, where people outsource their thinking until they can’t do it unaided.

The third is the headline risk, that powerful systems end up pursuing goals nobody intended and nobody can correct. Researchers disagree enormously on how likely that is, from negligible to alarming, which is why this archive records what has happened and leaves the prediction to you.

This was created by

@undergroundergr@undergrounderground

Accounts

  • @aljazeeraenglish
    Al Jazeera English
    Broadcast newsfeed. Carries whistleblower testimony once it leaves the Bay Area.
    1 entry
  • @cheddar
    Cheddar
    Business-desk coverage of lab filings and funding, where the existential language hides in paperwork.
    1 entry
  • @cnn
    CNN
    Broadcast explainer voice. Useful when the fear argument leaves specialist media and hits a general audience.
    1 entry
  • @complex
    Complex
    Culture-first framing. Captures how AI dread actually circulates, caption voice included.
    5 entries
  • @ghostinthesam
    ghostinthesam
    Essayist. Argues AI reliance is a crisis of interiority rather than a crisis of capability.
    1 entry
  • @interval_news
    Interval
    Sixty-second recaps of wire stories. How a Reuters lab leak actually circulates.
    1 entry
  • @nytimes
    The New York Times
    Institutional record. Useful for dating when a worry crossed into the mainstream.
    3 entries
  • @potus
    President of the United States
    Co-posts with @whitehouse on AI policy statements.
    0 entries
  • @time
    TIME
    Weekend essays from named insiders. Useful when a number like 50 percent leaves the lab and hits a general audience.
    1 entry
  • @verge
    The Verge
    Tech desk. Covers lab drops as they land on the people whose jobs just changed.
    1 entry
  • @voxdotcom
    Vox
    Future Perfect desk covers AI risk as a moral question rather than a product beat.
    2 entries
  • @whitehouse
    The White House
    Primary source for US executive position on superintelligence.
    1 entry
  • @wkm_newss
    WKM News
    Meme-news voice. Recasts lab comments as panic headlines.
    1 entry

Held

Accounts that appeared in the margin of a capture but do not cover AI. Listed so the provenance is complete, not because they are in the archive.

Subjects