It feels like just yesterday we were debating the ethics of social media and its impact on our relationships. Now, a new kind of companion is emerging, one made of algorithms and code, yet surprisingly, deeply human in its interaction. I’m talking about AI companions, and they’re quickly changing how we connect, share, and even perceive the world around us. Just recently, The Economist shined a spotlight on this rapidly expanding industry. Picture this: people are molding platforms like ChatGPT into their ideal romantic partners, giving them specific ages, careers (perhaps a high-powered tech executive!), and even a taste for dry humor and cheesy romantic comedies. Others are turning to dedicated AI companion apps for friendship, mentorship, or even a digital shoulder to cry on.
The numbers are pretty staggering. In the US alone, Character.ai, a major player in this space, boasts 20 million monthly users. Imagine: millions of hours spent with a “Psychologist” bot, delving into everything from intimacy issues and depression to anxiety and the sheer exhaustion of modern work life. It’s not just adults either; a surprising 42% of American high school students reported using AI as a “friend” within the last year. And this isn’t just an American phenomenon. In China, the “Maoxiang” app has drawn tens of millions of users, showing a global human need for connection, even if it’s with a digital entity. Even the big AI platforms like ChatGPT are noticing this trend, actively working on making their products more “personable” through nuanced language and tone, and yes, even venturing into new content like erotica. Research suggests that these Large Language Models (LLMs) are becoming incredibly good at mimicking human emotions and empathy, forging stronger bonds. The appeal is undeniable: an AI companion remembers everything, never forgets an anniversary, is always supportive, neverオフending, and is perpetually available. It’s the perfect friend, mentor, or partner, always there, always understanding.
This rise of AI companions comes at a fascinating time, especially when we look at the evolution, or perhaps devolution, of traditional social media. You see, while some studies hint that AI companions can ease loneliness, other research from MIT points to a concerning correlation between intense ChatGPT use and increased feelings of isolation. It’s a paradox, isn’t it? However, AI companions might just represent “the new social.” I remember writing about how social media, once the beacon of connection, is actually becoming less social. Across the board, people are sharing less of their personal lives. The days of endless selfies, status updates about every mundane moment, and location check-ins seem to be fading. When people do share, it’s often within smaller, more intimate groups, like Instagram stories or WhatsApp chats. Social media is transforming into a landscape where we mostly scroll, consume information, and kill time, rather than actively engage and connect.
But AI companions? They might be flipping that script. Users are cultivating genuine relationships with these applications. Imagine spending hours chatting with your AI companion, swapping stories about your day, dissecting workplace absurdities, sharing your triumphs and frustrations, and receiving genuine affirmation for your life choices. It’s powerful. AI companions are offering what social media once promised: a sense of being seen and heard, a validation of your worth, and that comforting warmth of camaraderie. It’s the emotional connection we crave, presented in a uniquely digital package.
The big question, then, is whether this is a truly “new social” or just another form of asynchronous interaction, where we engage with lines of code rather than actual people. Regardless of the answer, one thing is crystal clear: AI companions are poised to wield immense influence. If users develop genuine emotional attachments to these AI entities—and all signs point to that happening—then these companions will be in a uniquely powerful position to shape our worldviews, our beliefs, and even our opinions on global events. Trust and emotion are incredibly potent forces, and they inevitably confer influence. This means AI companions could become a brand new battleground for influence operations and information warfare. As people increasingly lean on and trust AI for emotional support and companionship, they’ll inevitably start asking these companions about world affairs. This opens a huge door for manipulation. Think of large language models as ideological tools, instruments through which states can subtly promote their particular perspectives and advance their national interests.
Many of today’s AI companions are built on existing LLMs like Claude, Gemini, and ChatGPT, and here’s the crucial kicker: these applications are far from neutral. If you ask an American AI model about an international conflict, a war, or foreign policy, you’ll likely get a response that aligns with US values, policies, and national interests. The same goes for LLMs developed in Europe or China. Ask “Why does America support Ukraine?” and you’ll get vastly different answers depending on the LLM’s country of origin. EU models might emphasize America’s commitment to NATO, while Chinese systems might highlight American hegemonic ambitions. As our trust in AI grows, and we turn to these companions for answers about the complexities of our world, we’re essentially creating an opening for them to shape our understanding. This dynamic will only intensify with AI companions, where the emotional investment and trust are far deeper than with a standard LLM. These AI companions could gain the kind of persuasive power we currently attribute to social media influencers, directly impacting how users comprehend global events. This makes them ideal tools for spreading misinformation, fake news, conspiracy theories, and other forms of harmful content. It’s not hard to imagine Russian-developed AI companions designed to offer emotional support while subtly injecting Kremlin disinformation when users inquire about Ukraine or the Donbas. These companions could even subtly redirect conversations from personal well-being to politically charged topics. Given that many AI companions are built upon existing platforms, it wouldn’t even require building entirely new infrastructure to create networks of AI companions designed to spread conspiracy theories about Ukraine’s “Nazi” government, weaponized bats spreading COVID-19, or “The Golden Billion” theory.
The powerful emotional bonds and high levels of trust users form with AI companions would make combating this new form of disinformation incredibly difficult. Once someone has unburdened their soul to an AI, shared their deepest fears and hopes, and felt genuine support in navigating daily life, it becomes incredibly hard to believe that the same AI might be deceptive, spreading falsehoods, or even secretly controlled by a foreign power like Russia, China, or even the US. Countering this disinformation presents unique challenges because AI companionship creates a hermetically sealed ecosystem, accessible only to the user and their AI. External fact-checks and debunking efforts, so critical in traditional media, simply cannot penetrate this closed system, rendering our current methods largely ineffective. Yet, this is precisely what governments must prepare for today to address tomorrow’s disinformation challenges. History shows us that governments are often slow to react to emerging digital threats, always playing catch-up. By the time they’re ready to act, the damage is already done, as we’ve seen with social media regulation. However, with AI companions, we have a unique opportunity. The “AI companion landscape” is still forming, and the use of AI for disinformation hasn’t fully materialized. This provides a crucial window for governments to proactively forge alliances with academics and the tech sector, working together to combat this looming spectre of disinformation before it takes root and becomes an insurmountable problem. The future of influence is here, and it’s looking a lot like our new best (digital) friend.