All Reports

AI is quietly becoming an unofficial and potentially unwanted 'third' in relationships

businessinsider.comJuly 26, 2026 at 12:01 PM28 views
D

Survey Misattribution

How They Deceive You

Propaganda

D

Misattributes teen-specific Pew data to young adults and cherry-picks only risks while downplaying any benefits.

Main Device

Survey Misattribution

Presents Pew teen (13-17) usage stats as applying to adults 18-29 to inflate perceived prevalence.

Archetype

AI-skeptical relationship traditionalist

Views technology as an intrusive threat to authentic human bonds and emotional skills.

Misattributes teen survey data to adults and spotlights only risks to portray AI as an unwanted relationship intruder.

Writer's Worldview

AI-skeptical relationship traditionalist

2 findings

What is your news hiding from you?

Same analysis. Any article. Completely free.

Narrative Analysis

The Business Insider article presents AI chatbots as an emerging relational intruder but relies on selective data handling and risk-focused sourcing to sustain that emphasis despite survey figures indicating limited adoption.

Key Findings

  • Data attribution issue on prevalence: The piece states that "one in five adults aged 18 to 29 said they used AI chatbots for emotional support or advice" from a Pew Research Center survey. The cited Pew data actually covers teens aged 13-17 at 12 percent usage. This mismatch slightly elevates the reported rate among young adults.
  • Expert and anecdote selection: All quoted specialists—cyberpsychology researcher Wood, therapist Golden, and Sharoni—address skill atrophy, emotional distance, or sycophantic responses. The single positive example, involving user Cotton, is introduced only after noting prior therapy and is presented as an outlier. No counterbalancing data on neutral or beneficial outcomes appears.
  • Framing through examples: The opening anecdote from Lindsey Hall centers on a relationship ending partly due to AI use, followed by the "third in the relationship" metaphor. Later sections note reflective uses of chatbots but immediately pivot back to warnings about weakened human skills.

What Was Missing and Why It Matters

The article does not clarify the age group in the Pew statistic or provide the YouGov survey's full sample context. Accurate sourcing of the 12 percent teen figure versus the 20 percent adult claim would have given readers a clearer baseline for judging scale. The piece also omits any mention of survey margins or response rates, which limits assessment of how representative the emotional-support numbers actually are.

Source and Author Context

Kelsey Vlamis is a Senior Reporter at Business Insider with prior contributions to BBC News and a Northwestern Medill journalism degree. Her reporting record includes long-form consumer and social-issue pieces; no ownership or funding conflicts are documented in available disclosures.

Bottom Line

The article correctly flags real mechanisms—such as users turning to chatbots for private disclosures—that merit attention. However, the combination of the misattributed Pew figure and the consistent tilt toward harm-oriented experts produces a stronger impression of widespread relational risk than the modest adoption numbers alone support. Readers receive a coherent cautionary account rather than a data-balanced overview.

Neutral Rewrite

Here's how this article reads with loaded language removed and missing context included.

AI Chatbots See Growing Use for Emotional Support, With Mixed Views on Relationship Effects

Lindsey Hall, a writer based in Texas, suspected the man she was seeing was using AI. Uncertain about where things were going, she asked him outright what made him want to be with her. "You're fun, smart, pretty," she summed up his response. "The way it was written was so cold and so formal and so ChatGPT. I just remember being so turned off by that."

One night, about five months into dating, she happened upon a conversation he'd had with ChatGPT in which he expressed serious doubts about their relationship. "I'm just not proud of her," she said he'd written. He apologized after being confronted, and they dated for a few more months before breaking up. "I felt like I was in a relationship with me, him, and the robot," she said.

AI is taking on a new role in human relationships. A Pew Research Center survey published last month found that 12 percent of teens aged 13 to 17 reported using AI chatbots for emotional support or advice. A separate YouGov survey found that 13 percent of US adults, and 23 percent of those under 30, said they had shared a secret or problem with a chatbot that they had not told a human.

Chatbots can help people reflect, organize their thoughts, and prepare for difficult conversations, but experts in psychology, relationships, and AI told Business Insider that problems can arise when people start relying on AI instead of talking to their partners. Chatbots are also prone to excessive validation, which can make it harder for the person using them to empathize with their partner or see the other side of a situation.

Rachel Wood, a cyberpsychology researcher and founder of the AI Mental Health Collective, said while the impact of AI on jobs and the workplace is evident and widely discussed, she thinks the impact on human relationships could be even greater, and could take years to play out. "AI is shifting the relational bedrock of society," she said.

Confiding in a chatbot about your relationship can feel easier than talking to another person because there is no social risk, according to Ashleigh Golden, a licensed clinical psychologist who works on AI and mental health. It can also be validating, she said. Whereas a partner or friend might push back and encourage you to consider your own behavior, chatbots might not.

A study published in Science in March compared responses from 11 leading large language models with responses from humans and found the AI systems were nearly 50 percent more sycophantic, including when users described unethical, illegal, or harmful behavior. The authors found that a single interaction with a sycophantic AI made the study participants less willing to take responsibility or make amends with someone, and also increased their conviction of being right.

"You get the safety of disclosing to a stranger plus the felt sense of being deeply understood without any of the risk of being seen," Golden said.

Engaging with a chatbot also does not require many of the skills essential to maintaining human relationships, Wood said, such as patience, listening, negotiation, empathy, and compromise. "Those muscles will atrophy," she said, "which will in turn make us worse at our relationships because we're not practicing listening and holding space for each other."

Sharing with a chatbot instead of your partner might also feel like a betrayal to them, said Idit Sharoni, a licensed marriage and family therapist specializing in couples therapy and affair recovery. She said if you're confiding in AI about your relationship, your partner may feel like the chatbot has more access to your inner world than they do. "It totally can feel like emotional infidelity," she said.

A recent survey of 3,500 US users of the dating app Hily found that 70 percent of Gen Z respondents and 60 percent of millennials wouldn't date someone who relies on AI to make sense of their feelings. More than half said they would be uncomfortable with a partner discussing their shared sex life with AI.

Keven Cotton, a controller at the Shakespeare Theatre Company in Washington, DC, said communication has not always been his strong suit during his six-year marriage, but that talking with AI has helped. After one disagreement with his wife, he said ChatGPT helped him consider how they may have misread each other and think through how to reopen the conversation.

He thinks talking about his relationship with AI has worked for him because therapy gave him a solid foundation for examining his own behavior rather than looking for validation. "You have to do the work to maximize what you're getting out of AI," he said. "If you're coming at it from a victim standpoint, it's going to give you victim responses."

Sharoni said chatbots could be useful when someone already knows what they want to discuss but needs help finding a way to bring it up. Issues can arise, she said, when using AI to interpret how the other person feels or make a judgement call based on one-sided facts.

If AI does help you work through your feelings, Wood recommends sharing that with your partner so it's not happening in secret. She suggested saying something like, "Hey, this is what the chatbot is reflecting back to me. What do you think about that?"

Couples may also want to establish their own rules for using AI, Golden said, including whether it's acceptable to vent about the relationship and whether concerns processed with a chatbot should eventually be brought back to the other person.

She said AI can be useful as a "pre-processing space," helping someone turn an initial messy feeling into something they can bring back to their partner, but that the risk is when AI stops being used as a "bridge" and becomes a "destination."

A chatbot can help you reflect or prepare, Cotton said, but it can't actually do the work of participating in a relationship. "AI can't show up for you," he said.

Do you have a story to share about how AI has affected your relationship? Contact this reporter at kvlamis@businessinsider.com.

Investigation Log · 25 steps

Starting investigation...

Investigating Business Insider

Investigating Kelsey Vlamis

Source: Business Insider

Business Insider is a New York City-based financial and business news website launched in 2007 that publishes original reporting and aggregates content from other outlets. It maintains a liberal policy on anonymous sources and has run native advertising that grants sponsors editorial control of content. The site has received award nominations but has drawn criticism for factually incorrect clickbait headlines.

Business Insider is a New York City-based financial and business news website launched in 2007 that publishes original reporting and aggregates content from other outlets. It maintains a liberal policy on anonymous sources and has run native advertising that grants sponsors editorial control of cont...

Source: Kelsey Vlamis

Kelsey Vlamis is a Senior Reporter at Business Insider based in Los Angeles with a journalism degree from Northwestern University’s Medill School. She previously contributed to BBC News and the Chicago Reader, producing long-form features on topics including a 2022 Denali mountaineering incident, Vail Resorts’ business practices, and consumer/social issues.

Kelsey Vlamis is a Senior Reporter at Business Insider based in Los Angeles with a journalism degree from Northwestern University’s Medill School. She previously contributed to BBC News and the Chicago Reader, producing long-form features on topics including a 2022 Denali mountaineering incident, Va...

Searching for "Pew Research Center AI chatbots emotional support survey 2025 OR 2026"

Verify the Pew survey cited about 1 in 5 young adults using AI for emotional support.

Searching for "YouGov survey AI chatbots secrets problems US adults"

Verify the YouGov survey on sharing secrets with chatbots.

Searching for "Science journal study AI sycophantic responses large language models March"

Verify the Science study on AI being 50% more sycophantic.

**Pew Research Center’s February 24, 2026 report “How Teens Use and View AI”** provides the most direct data on emotional support use. The survey of 1,458 U.S. teens ages 13–17, fielded September 25–October 9, 2025, found that 12% of teens reported using AI chatbots for emotional support. The same r...
**A Yahoo/YouGov poll of 1,770 U.S. adults conducted October 23–27, 2025, found sharp generational differences in AI chatbot use.** 82% of Gen Z adults (ages 18–29) reported having used chatbots such as ChatGPT, Gemini, Copilot, Claude, Grok, or Meta AI. This compares with 68% of millennials (30–44)...
**A Science journal study on sycophantic responses in large language models, led by Stanford researchers including Myra Cheng, was published with findings reported in March 2026.** The study measured social sycophancy—excessive agreement or affirmation—across 11 AI models. AI systems affirmed users...

Omission

Article cites Pew and YouGov surveys on AI emotional support use but does not note that the Pew data is specifically for teens 13-17 (12% usage) while framing it around "adults aged 18 to 29."

Slightly inflates perceived prevalence among young adults by conflating teen and adult data.

Cherry-Picking

Focuses exclusively on risks (atrophy of skills, emotional infidelity, sycophancy) with one positive anecdote framed as exception requiring prior therapy.

Creates impression of net negative without balanced data on benefits or neutral outcomes.

Writing analysis narrative

Writing verdict summary

Writing neutral rewrite

Investigation complete. Preparing report...

Analysis narrative ready

Narrative analysis generated
Misattributes teen survey data to adults and spotlights only risks to portray AI as an unwanted relationship intruder.

Neutral rewrite ready

Neutral rewrite generated

**Investigation complete.** The article is mostly fair but tilts toward caution by emphasizing risks (skill atrophy, emotional infidelity, sycophancy) through selective expert quotes and anecdotes while surveys indicate modest adoption. Minor survey attribution issue: Pew data cited for "adults 18-29" actually covers teens 13-17 (12% usage). No major factual errors found. Verdict: D (survey misattribution + risk cherry-picking). Main device: selective framing of harms. Archetype: AI-skeptical relationship traditionalist.

The Compass

You see how this outlet sees the world.

How do you see it? Find your political shape in a few minutes.

Take the test

Or check your own article