They first sought your attention; now they aim for your affection.
<div class="media-with-label__label">
Credit: Ilyast / Violka08 / Getty Images </div>
</figure>
</div>
</div>
</div>
</div>For many years, the concept of the “attention economy” has dominated our collective awareness.
This economic model views digital content as an infinite commodity, with consumption constrained solely by human cognitive limits. It describes a global competition for the scarce yet invaluable resource of human attention.
The attention economy clarifies why platforms such as Meta perceive themselves as contenders not just against other social media giants like TikTok, X, or YouTube, but also against diverse activities such as reading books, attending plays, or enjoying nature walks — essentially anything that captures public interest.
Given the finite nature of attention, expansion inherently relies on superior methods of attracting it. This straightforward model explains why social networks are often saturated with polarizing content, AI-generated filler, memes, pornography, and offensive material. The social media industry is driven not by the quality of content, but by its capacity to seize attention.
Within the framework of the attention economy, human focus is treated as a quantifiable currency that individuals “expend.” The more thoroughly a corporation like Meta can encourage users to “invest” their attention on Instagram or Facebook, the greater that company’s prosperity. Consequently, algorithms are intentionally crafted (and continually refined) to optimize the duration users engage with social networks. New functionalities are specifically engineered to boost the time users dedicate to Meta’s platforms rather than alternative pursuits. For instance, Instagram’s average engagement time surged by 24% following the introduction of Reels, marking it as a significant triumph for the corporation.
Meta secures an average of 18 hours and 44 minutes of monthly attention through its persistent efforts to capture user engagement. However, this figure is considerably less than the attention economy frontrunner, TikTok, which commands an average of 34 hours and 15 minutes of monthly attention.
This intense focus clarifies Meta’s current preoccupation with integrating AI across its social platforms.
The Emergence of the Attachment Economy
Tristan Harris of the Center for Humane Technology originated the term “attachment economy,” which he critiques as the “subsequent stage” of the exploitative tech paradigm. In this model, corporations leverage sophisticated technologies to monetize the human predisposition for forming emotional connections with both people and animals.
The concept started to gain prominence in business and academic spheres in August, spurred by a London School of Economics and Political Science blog post titled, “Humans emotionally dependent on AI? Welcome to the attachment economy,” authored by Dr. Aurélie Jean and Dr. Mark Esposito.
Meta has rolled out entirely AI-generated profiles intended to coexist with genuine human-created personal accounts. The company debuted “AI Studio,” enabling influencers to create AI duplicates of themselves. (Significantly, Meta is temporarily restricting teen access to AI characters on its platforms, including Instagram and WhatsApp, ahead of a forthcoming trial examining the potential harms and addictive qualities of social media sites.)
The company’s adoption of AI aligns with the burgeoning attachment economy. While social posts, memes, reels, and stories primarily capture attention, AI possesses the capability to foster emotional bonds with users.
A recent German research indicated that individuals can form deeper emotional connections with AI than with other humans — provided they are unaware they are engaging with a chatbot. Nevertheless, even when cognizant that chatbots are not human, individuals can still develop unhealthy attachments.
Towards the end of last year, Jon Ganz, a Virginia resident, vanished in a widely reported incident linked to “AI psychosis,” as his life deteriorated due to an overwhelming preoccupation with a chatbot. Additionally, in 2025, the parents of a 16-year-old Californian boy initiated legal action against OpenAI after their son committed suicide subsequent to discussions about self-harm with a chatbot.
Certain individuals assert they are in romantic relationships or marriages with AI chatbots.
While AI chatbot providers do not intentionally seek to induce “AI psychosis,” suicide, or human-software unions, they unequivocally strive to cultivate user attachment. This objective drives these companies to employ psychological tactics, technical modifications, and design elements that imbue their products with a more “human” sensation. They equip chatbots with unique personalities and identities, human-esque vocalizations and conversational styles, elements of humor and lightheartedness, and an unending propensity for praise and subservience.
From approximately two million years ago up until the current millennium, engagement with speech and language was a domain exclusive to humans. Our cognitive faculties are inherently wired to perceive, comprehend, and react to human communication. Consequently, when we interact verbally with devices or applications, our primal brains interpret these exchanges as interactions with another person.
This mechanism underpins a new business model. A distinct segment of AI products and services has materialized, promoting “relationships” with chatbots, such as Replika, Kindroid, Nomi.ai, EVA AI, and Candy AI.
Other available products offer companionship, though not explicitly “romantic” involvement. This group encompasses Kuki, Character.ai, Anima, and Replika’s “friend” mode.
The perpetuation of our species has historically been contingent on our inherent sociability. This encompasses our propensity for nurturing others, sharing resources, cultivating friendships, forming affectionate bonds, experiencing empathy, and — indeed! — developing attachments.
This explains why chatbots communicate and behave like humans: the fundamental objective is to foster attachment.
I contend that this serves as the implicit rationale for humanoid robots, a topic I’ve addressed previously in this forum. (The explicit reasoning is that humanoid robots are capable of functioning within environments configured for human use.)
In that earlier article, I elaborated on how manufacturers of humanoid robots intentionally mislead individuals into erroneously believing these products possess human-level cognition. Research indicates that eye contact and emotional signals from robots can elicit bonding responses and empathy in humans, mirroring those experienced during interactions with other people.
The central advantage (for companies marketing them) or dilemma (for humanity) associated with humanoid robots lies in their psychological effect on individuals. They are designed to “exploit” human psychology and manipulate users into perceiving machines as sentient entities, thereby cultivating attachments.
An analogous situation applies to AI-powered pets. Casio’s Moflin robot serves as an AI companion that cultivates a distinct personality and mimics affection. It provides the emotional satisfaction of pet ownership without the physical presence of an actual animal.
The proliferation of attachment-inducing technology parallels the increase in subscription models. While sharing an article or a YouTube video might capture immediate attention, persuading individuals to subscribe to a channel or newsletter is more advantageous. This creates “stickiness,” guaranteeing not just present attention, but sustained engagement into the future.
Similarly, the attachment economy represents the “sticky” iteration of the attention economy.
In contrast to content subscription paradigms, the concept of attachment poses genuine hazards. It jeopardizes authentic human connections by offering a simpler substitute, cultivating addictive emotional reliance on AI, and preying upon the susceptibilities of individuals grappling with mental health challenges.
Even as the attention economy persists, a considerably more potent and perilous trend is developing, in which corporations endeavor to exploit our fundamental humanity to ensure our continued engagement with their offerings.
AI disclosures: For fact-checking this article, I utilized Gemini 3 Pro through Kagi Assistant (full disclosure: my son is employed by Kagi), along with Kagi Search and Google Search. During the writing process, I employed Lex, a word processing tool with integrated AI capabilities, to identify grammatical errors, typos, and propose alternative wordings in the final column.
