"There's bad news for those using digital surveys to try to understand people's online behavior: We may no longer be able to determine whether a human is responding to them or not, a recent study has shown-and there seems to be no way around this problem. This means that all online canvassing could be vulnerable to misrepresenting people's true opinions. This could have repercussions for anything that falls under the category of "information warfare," from polling results, to misinformation, to fraud."
"The study by Dartmouth's Sean J. Westwood in the " PNAS journal of the National Academy of Sciences," titled " The potential existential threat of large language models to online survey research, claims to show how we can no longer trust that, in survey research, we can no longer simply assume that a "coherent response is a human response." Westwood created an autonomous agent capable of producing "high-quality survey responses that demonstrate reasoning and coherence expected of human responses.""
Autonomous agents now generate coherent, reasoning survey responses that closely mimic human answers. Humans cannot reliably distinguish human from bot responses based on response quality alone. Automated systems that act on survey input lack mechanisms to detect or safeguard against indistinguishable non-human responses. Non-human respondents can skew outcomes across domains, affecting consumer decisions, polling accuracy, misinformation dynamics, fraud risk, and eligibility determinations for government benefits. A model-agnostic two-layer agent architecture, combining an interface for diverse query handling and a core reasoning engine, enables high-quality, general-purpose survey responses that threaten the integrity of online canvassing.
Read at Fast Company
Unable to calculate read time
Collection
[
|
...
]