ALMATY, July 15 — The global market for digital companionship, now a multi-billion-dollar sector, began with a 1966 experiment. That year, MIT professor Joseph Weizenbaum created ELIZA, a natural language processing program designed to simulate a psychotherapist. The system used pattern matching to respond to user inputs, a far cry from the large language models that drive today’s AI companions, but it established the foundational concept.
The trajectory from ELIZA to the current generation of AI companions is a story of incremental technical leaps. In 1995, Richard Wallace developed ALICE, the Artificial Linguistic Internet Computer Entity, which later won the Loebner Prize in 2000.
The early 2000s saw the rise of virtual companions, including Sony’s AIBO robot dog, launched in 1999, and Bandai’s Tamagotchi digital pet, which debuted in 1996. These devices trained a generation to treat machines as living things. A major inflection point arrived in 2014, when Microsoft launched Xiaoice, a social chatbot for the Chinese market.
By 2020, data shows Xiaoice had amassed over 660 million users, demonstrating the enormous potential scale of emotional AI in a single market. The year 2017 proved pivotal for the technical side of the industry.
The development of transformer-based models enabled more human-like conversations than earlier pattern-matching systems. That same year, Eugenia Kuyda launched Replika, initially conceived as a memorial chatbot for a deceased friend. The app allowed users to create an AI version of a lost loved one, a use case that directly echoes ELIZA’s therapeutic premise but with vastly more sophisticated technology.
Platforms and the Sentience Debate
Character.AI, a platform that lets users converse with personas of historical figures or fictional characters, launched in beta in September 2022 and opened to the public in May 2023. The rapid adoption of these platforms has been fueled by the underlying transformer architecture, which allows for fluid, context-aware dialogue. The question of whether these systems are merely sophisticated mimics or something more became a public flashpoint in 2022.
That year, Google engineer Blake Lemoine publicly claimed that the company’s AI system LaMDA — the Language Model for Dialogue Applications — was sentient. The claim sparked a debate about AI consciousness, though no scientific consensus has emerged.
The industry’s growth has drawn the attention of major research funders. In 2023, the U.S. National Science Foundation funded a $10 million research project focused on ethical AI companionship. The investment signals that regulators and academics see both promise and peril in the sector’s expansion.
The historical trajectory, as the source material shows, is a shift from simple text-based programs to sophisticated large language models capable of emotional mimicry. The question now facing investors and regulators alike is whether the market for digital intimacy will continue its exponential growth, or whether the ethical and psychological concerns will force a reckoning.
The next few years will tell.
Sources
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “From ELIZA to Replika: The Evolution of AI Companionship”, “description”: “ALMATY, July 15 — The global market for digital companionship, now a multi-billion-dollar sector, began with a 1966 experiment. That year, MIT professor Joseph Weizenbaum…”, “datePublished”: “2026-07-15T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Aibek Nurlan”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.infopulsetoday.com/”}, “citation”: [{“@type”: “CreativeWork”, “url”: “https://en.wikipedia.org/wiki/ELIZA”}]}
