Like it or Hate it, 64% of Teens Use AI Chatbots

Pew Research Center released survey data this week showing that 64% of U.S. teens ages 13-17 use AI chatbots, with roughly 30% using them daily and 16% using them multiple times per day or almost constantly. ChatGPT dominates usage at 59%, more than double the next most common platforms—Gemini (23%) and Meta AI (20%).

The data also documents continued social media saturation: YouTube reaches nearly all teens (roughly 90%), while TikTok and Instagram are used by about 60% or more. Approximately one-in-five teens report being on TikTok and YouTube "almost constantly," and 40% of teens say they're online almost constantly overall—up from 24% a decade ago.

These are descriptive statistics about behavior patterns. What they don't tell us: whether this level of AI chatbot and social media engagement is developmentally appropriate, educationally beneficial, psychologically harmful, or simply the new normal that will seem unremarkable in five years.

The Demographic Patterns That Raise Questions

Pew's data reveals significant disparities across race, ethnicity, age, and household income. Black and Hispanic teens are more likely than White teens to use chatbots (roughly 70% versus 58%), use them daily (35% and 33% versus 22%), and report being online almost constantly (55% and 52% versus 27%).

Lower- and middle-income teens are more likely than higher-income teens to use TikTok, Facebook, and be online almost constantly. They're also more likely to use Character.ai (14% versus 7%), while higher-income teens are more likely to use ChatGPT (62% versus 52%).

These patterns could reflect multiple dynamics: differential access to devices and internet connectivity, varying parental monitoring practices, different educational contexts where AI tools are introduced, or platform-specific targeting and adoption trends. Pew's survey documents the disparities but doesn't explain causation or consequences.

The question these numbers raise: are we seeing technology adoption patterns that reflect and potentially amplify existing socioeconomic inequalities, or are we seeing organic differences in how different communities integrate new tools?

What "Almost Constantly" Actually Means

Forty percent of teens report being online almost constantly. One-in-five say the same about TikTok and YouTube specifically. These are self-reported measures without objective screen time data, which means they capture teens' subjective experience of connectivity rather than precise usage metrics.

"Almost constantly" could mean checking platforms every few minutes, having them open in the background while doing other activities, or feeling psychologically connected even when not actively using devices. The phrase conflates different types of engagement that may have different implications.

Pew notes this represents an increase from 24% a decade ago, though it's a slight dip from last year. Whether this reflects stabilization, measurement noise, or the beginning of a trend reversal isn't clear from a single data point. What is clear: a substantial percentage of teens experience their relationship with online platforms as continuous rather than episodic.

That could be concerning if it displaces other developmental activities, beneficial if it enables connection and learning, or neutral if teens are simply using digital tools the way previous generations used telephones and television—constantly available but integrated into broader life patterns rather than replacing them.

ChatGPT Dominance and What It Signals

ChatGPT's 59% usage rate among teens significantly exceeds other chatbots. This could reflect several factors: earlier market entry giving it name recognition, integration into educational contexts, superior capabilities, or simply being the default "AI chatbot" that people try first.

The gap between ChatGPT and competitors (Gemini at 23%, Meta AI at 20%) is substantial enough to suggest winner-take-most dynamics rather than a diverse, competitive chatbot market. That concentration matters if ChatGPT shapes how teens conceptualize AI capabilities, what they expect from conversational interfaces, or how they approach information-seeking and problem-solving.

Pew doesn't ask what teens use chatbots for—homework help, creative projects, entertainment, emotional support, general curiosity. Usage statistics without context about application can't distinguish between students getting help with algebra, people generating creative fiction, or teens seeking advice on personal problems from AI systems designed for different purposes.

The demographic variations in chatbot preference are also interesting. Lower-income teens favor Character.ai, which focuses on fictional character conversations and roleplay. Higher-income teens favor ChatGPT, which is positioned as a general-purpose productivity tool. This could reflect different use cases, different discovery paths, or different perceptions about what AI chatbots are for.

The Social Media Baseline That Isn't Changing

YouTube maintains near-universal reach (roughly 90%), which has remained stable. TikTok and Instagram usage is also relatively stable from recent years, though both see daily usage rates around 60% and 55% respectively. Facebook continues declining among teens (31% usage, down from 71% in 2014-15), which is unsurprising given its demographic shift toward older users.

What stands out is how entrenched these platforms have become. The shares using YouTube, TikTok, and Instagram "have stayed relatively stable in recent years," meaning teens aren't abandoning these platforms despite ongoing concerns about mental health impacts, content moderation failures, and privacy issues.

This stability could indicate satisfaction with existing platforms, lack of viable alternatives, network effects making switching costs prohibitive, or simply that teens compartmentalize platform-specific controversies from daily usage decisions. The data doesn't distinguish between active preference and path-dependent continuation.

Pew notes that WhatsApp grew from 17% teen usage in 2022 to 24% in 2025, which is notable given it's primarily a messaging app rather than a social media platform. This could reflect teens seeking more private communication channels, international adoption patterns filtering into U.S. usage, or simply diversification of communication tools.

What the Research Doesn't Measure

Pew's survey captures self-reported platform usage and frequency. It doesn't measure outcomes: academic performance, mental health indicators, social relationship quality, information literacy, or any other metric that would help assess whether these usage patterns are beneficial, harmful, or neutral.

The researchers also don't ask about context: Are teens using chatbots for homework because teachers assign it, because it's more effective than traditional study methods, or because it's easier than learning material directly? Are they on TikTok "almost constantly" because the algorithm is optimized for engagement, because their social lives depend on it, or because they genuinely find it valuable?

Without outcome data or contextual understanding, these statistics describe behavior without explaining why it's happening or what it means for teen development. That's a limitation of survey methodology, not a criticism of this specific research. But it's important for interpreting what these numbers actually tell us.

The Income and Race Disparities Nobody's Addressing

The demographic patterns Pew documents raise questions that extend beyond technology adoption. Why are Black and Hispanic teens roughly twice as likely as White teens to report being online almost constantly? Why are lower-income teens more likely to use certain platforms while higher-income teens favor others?

Possible explanations include: differences in parental monitoring and household rules, variations in school technology policies, different social norms around screen time across communities, disparities in access to alternative activities (extracurriculars, outdoor recreation), or platform-specific marketing and feature development targeting different demographics.

These aren't mutually exclusive, and the survey data can't adjudicate between them. What it does show is that technology adoption and usage intensity correlate with socioeconomic factors in ways that could either mitigate or amplify existing inequalities depending on whether digital engagement produces positive or negative outcomes.

If chatbot usage improves educational outcomes, higher-income teens having greater access to ChatGPT (62% versus 52%) while lower-income teens use Character.ai more (14% versus 7%) could reflect and reinforce achievement gaps. If constant connectivity undermines wellbeing, higher rates among Black, Hispanic, and lower-income teens are concerning for equity reasons.

What This Data Actually Shows

Pew's research documents that AI chatbots have rapidly achieved majority adoption among U.S. teens within a relatively short timeframe, with usage patterns showing demographic variations that correlate with existing socioeconomic divides. It confirms that social media platforms remain central to teen life, with substantial minorities reporting near-constant connectivity.

What it doesn't show: whether any of this matters. Whether chatbot usage supports learning or enables plagiarism. Whether constant connectivity facilitates social connection or displaces face-to-face relationships. Whether platform engagement reflects genuine value or exploitation of psychological vulnerabilities.

These are empirical questions that require outcome research, not just usage statistics. The honest assessment: we're documenting rapid adoption of new technologies among adolescents without understanding developmental implications. That's concerning if you think precaution should precede mass adoption. It's inevitable if you think technologies diffuse faster than research can evaluate them.

Either way, the data shows teens are integrating AI chatbots into daily life at scale, building habits and expectations around tools whose long-term effects remain uncertain. Whether that's a problem we should address or simply reality we need to understand is a judgment call these statistics can't make for us.

If you're evaluating how emerging technologies affect user behavior and need research that goes beyond adoption metrics to understand actual impact, Winsome's team can help you design studies that answer strategic questions rather than just documenting trends.

Source: Pew Research Center, "Teens, Social Media and AI Chatbots 2025" by Michelle Faverio and Olivia Sidoti