In the AI era, we will always have to fight for the truth / Das Zeitalter der KI wird ein ständiger Kampf um die Wahrheit prägen (Jonas Schaible, 2026).
The problem is concrete: AI can be poisoned easily, as we found and now others have replicated. Evidence abounds in the collection of the weirdest images ever published in academic journals.
Meanwhile, the AI optimism narrative persists. A new Science editorial by Holden Thorp documents the gap between promise and practice: “AI in scientific publishing: Slower, worse, and more expensive.” The promises remain seductive –
how it will transform work, claiming that so little human effort will be required that humanity will enter an era of radical abundance, free from disease, drudgery, and danger, among other benefits
yet empirical reality suggests otherwise. Bergstrom et al show that “The unintended consequences of large language models as a labor-augmenting technology in science” actually cut against the optimistic case:
By allowing scientists to work more quickly, LLMs raise the opportunity cost of researcher time, creating incentives to refine papers less thoroughly before moving on. Enticing as it is to imagine that, by saving us time on mundane tasks, LLMs will provide us with more time to think deeply and develop projects completely, our results temper such hopes.
The implication is stark: fighting for truth may now require more labor, not less.