Report: “The Chatbot Optimisation Game: Can We Trust AI Web Searches?”

From The Guardian:

But what is pitched as a more convenient way of looking up information online has prompted scrutiny over how and where these chatbots select the information they provide. Looking into the sort of evidence that large language models (LLMs, the engines on which chatbots are built) find most convincing, three computer science researchers from the University of California, Berkeley, found current chatbots overrely on the superficial relevance of information. They tend to prioritise text that includes pertinent technical language or is stuffed with related keywords, while ignoring other features we would usually use to assess trustworthiness, such as the inclusion of scientific references or objective language free of personal bias.

[Clip]

Gaming chatbots is possible, then, but not straightforward. And while website owners and content creators have derived an evolving list of essential SEO dos and don’ts over the past couple of decades, no such clear set of rules exists for manipulating AI models. The term generative engine optimisation was only coined last year in an academic paper, whose authors concluded that using authoritative language (regardless of what is expressed or whether the information is correct) alongside references (even those that are incorrect or unrelated to what they’re being used to cite) could boost visibility in chatbot responses by up to 40%. But they stress these findings aren’t prescriptive, and identifying the exact rules governing chatbots is inherently tricky.

[Clip]

Yet an internet dominated by pliant chatbots throws up issues of a more existential kind. Ask a search engine a question, and it will return a long list of webpages. Most users will pick from the top few, but even those websites towards the bottom of the results will net some traffic. Chatbots, by contrast, only mention the four or five websites from which they crib their information as references to the side. That casts a big spotlight on the lucky few that are selected and leaves every other website that isn’t picked practically invisible, plummeting their traffic.

Learn More, Read the Complete Article (about 1770 words)

Filed under: Journal Articles, News, Patrons and Users