In 1971, Herbert Simon (Nobel ’78) published an essay on the “Attention Economy.” There is a famous saying, “Abundance of information leads to lack of attention.” He provided insight into how economic organizations (and people) need mechanisms to receive and process large amounts of information and then pass on only the relevant parts of that information. (Simon won the Nobel Prize for his “pioneering research on decision-making processes in economic organizations.”)
Yaqub Chaudhary and Jonnie Penn suggest that artificial intelligence could change the parameters of the balance between information and attention and instead lead to the so-called “intention economy.” They describe this outlook in “Beware the Intent Economy: Collecting and Commoditizing Intents through Large Language Models,” published December 30, 2024. Harvard Data Science Review In a paper titled “Grappling With the Generative AI Revolution”.”
From the abstract:
The rapid proliferation of large-scale language models (LLMs) opens the door to new markets for behavioral and psychological data that reveal intent. This brief article introduces some early features of emerging markets. We examine recent efforts by technology executives to position the capture, manipulation, and commodification of human intent as a viable extension parallel to the currently dominant attention economy. Since the 1990s, the scope of interest has been limited. We call this the intent economy of follow-up. We characterize it in two ways. First, competition between established technology players armed with the infrastructure and data capacity needed to compete for a first-mover advantage in an initially compelling new frontier of technology. Second, hitherto unattainable levels of explicit and implicit data that signal intent, i.e., combining (a) hyper-personalized manipulation through LLM-based flattery, ingratitude, and emotional penetration with (b) increasingly detailed classifications; This is to commercialize the generated signal. Online activities guided by natural language.
This new level of automated persuasion more broadly leverages the unique capabilities of LLM and generative AI to engage not only with what users want, but also, to quote Williams, “with what they want” (Williams, 2018, p. 122 ). ). A close reading of recent technical and critical literature (including unpublished papers on ArXiv) shows that these tools are already being explored to derive, infer, collect, record, understand, predict, and ultimately manipulate, falsify, and commercialize human plans. It shows. There are mundane purposes (e.g. choosing a hotel) and profound purposes (e.g. choosing a political candidate).
I confess that I am only partially convinced that the ‘intention economy’ is fundamentally new and different from the ‘attention economy’. Vance Packard’s classic book, hidden persuader– Written in 1957 about how our needs and desires can and are manipulated by corporations, the media and politicians. Chaudary and Penn wrote: “At the time of printing, the intent economy was more an aspiration than a reality.” But here’s the example they have in mind:
[A] A concrete example helps illustrate how the intent economy, a digital marketplace for commodified signals of ‘intent’, differs from today’s attention economy. Today, advertisers can tap into current user interests through things like real-time bidding. [RTB] Networks like Google AdSense) or in the future (e.g. buying next month’s advertising space on billboards or subway lines) LLM allows advertisers to use real-time (e.g. ‘Have you thought about seeing Spider-Man tonight?’) and possibly in the future (e.g. We diversify this market format by allowing you to bid on items (e.g. ‘You said you’re overworked, would you like to book tickets to that movie?’). Did we talk about it?’). If you’re reading these examples online, imagine that each example has been dynamically generated to match your personal behavioral tracking, psychological profile, and situational indicators. In the intent economy, LLMs can leverage user flows, politics, vocabulary, age, gender, preferences for flattery, etc., along with intermediary bidding at low cost to maximize the likelihood of achieving a given goal (e.g. movie tickets). if selling). Zuboff (2019) identifies these types of personal AI ‘assistants’ as the equivalent of ‘marketplace avatars’ who steer conversations across the services of platforms, advertisers, businesses and other third parties.
Simply put, imagine a persuasive message that is much more individualized in many senses. These messages may be based on quite a wide range of data about you, including where you live and work, your travel patterns, family status, past purchases, and past Internet searches. We may then compare your personal data with the personal data of other people to find people who are statistically similar to you. Depending on how you are categorized based on your personal data, the messages you receive will be expressed in language that is most likely to be of interest to you, based on how you have personally responded in the past and how others have responded statistically. I had a similar answer to yours. These messages can also be “dynamically adjusted.” This means that instead of receiving the same message over and over again, you receive an ever-changing series of messages.
Chaudary and Penn recognize that some of this sounds like better-targeted advertising, but they argue that it has “the potential to engage with and commodify a higher level of user intentionality than that seen in the attention economy.” Perhaps the bottom line is that AI tools are already starting to provide forward and backward interaction. Sometimes it comes through a series of advertisements, sometimes in the form of chatbots, and sometimes in the form of offers of medical advice or treatment. As these interactions increase, it is important to remember that AI is both a tool that you use and a tool that others can use to communicate with you. In either case, AI is not your friend that only has your best interests at heart.
The post Will AI bring about the ‘intention economy’? First appeared in Conversable Economist.
