Teens turn to AI for emotional support
October 13, 2025
Welcome back.
1. Teens turn to AI for emotional support
2. AI takes center stage in classrooms
3. SoftBank is building an AI warchest
CULTURE
Teens turn to AI for emotional support
Everybody needs someone to talk to.
More and more, young people are turning to AI for emotional connection and comfort. A report released last week from the Center for Democracy and Technology found that 19% of high school students surveyed have had or know someone who has a romantic relationship with an AI model, and 42% reported using it or knowing someone who has for companionship.
The survey falls in line with the results of a similar study conducted by Common Sense Media in July, which found that 72% of teens have used an AI companion at least once. It highlights that this use case is no longer fringe, but rather a “mainstream, normalized use for teens,” Robbie Torney, senior director of AI programs at Common Sense Media, told The Deep View.
And it makes sense why teens are seeking comfort from these models. Without the “friction associated with real relationships,” these platforms provide a judgment-free zone for young people to discuss their emotions, he said.
But these platforms pose significant risks, especially for young and developing minds, Torney said. One risk is the content itself, as these models are capable of producing harmful, biased or dangerous advice, he said. In some cases, these conversations have led to real-life harm, such as the lawsuit currently being brought against OpenAI alleging that ChatGPT is responsible for the death of a 16-year-old boy.
Some work is being done to corral the way that young people interact with these models. OpenAI announced in late September that it was implementing parental controls for ChatGPT, which automatically limit certain content for teen accounts and identify "acute distress” and signs of imminent danger. The company is also working on an age prediction system, and has removed the version of ChatGPT that made it into a sycophant.
However, OpenAI is only one model provider of many that young people have the option of turning to.
“The technology just isn't at a place where the promises of emotional support and the promises of mental health support are really matching with the reality of what's actually being provided,” said Torney.
Though OpenAI’s parental controls represent a solid step in protecting younger users from the dangers of these platforms, these measures are only a first step. Parental controls can easily be bypassed, for example, and don’t recognize when parents themselves are what are causing emotional distress. Along with stricter guardrails from the platforms themselves, regulation is necessary for preventing real harm. And while the EU has the AI Act, which explicitly recognizes children as vulnerable users requiring special protection and bans AI systems that manipulate minors or encourage dangerous behavior, the lack of regulation in the U.S. may only exacerbate this problem.
TOGETHER WITH OUTSKILL
This Halloween, skip the fake scares.
The real one? AI replacing your job.
Over 70% of people believe it’ll happen by 2025.
Here’s your way out 👇
Join the 2-Day LIVE AI Mastermind by Outskill — a hands-on crash course that turns you into an AI-powered professional who can build, earn, and automate with AI.
💥 Normally $395 — now FREE for Halloween!
⭐ Rated 9.8/10 on Trustpilot – an opportunity that makes you an AI Generalist who can build, solve & work on anything with AI, Instead of fearing it.
In just 16 hours (5 sessions), you’ll learn how to:
Build AI agents that save 20+ hours weekly
Master 10+ tools pros charge $150/hr to use
Launch your $10K+ AI consulting business
Automate 80% of your workload
🧠 Live sessions: Sat–Sun, 10 AM–7 PM EST
🎁 Get $5,000+ in AI bonuses — prompt bibles, AI monetization roadmap, and your personal toolkit.
STARTUPS
AI takes center stage in classrooms
AI is going back to school.
Campus, a college education startup backed by OpenAI’s Sam Altman, hired Jerome Pesenti as its head of technology, the company announced on Friday. Pesenti is the former AI vice president of Meta and the founder of a startup called Sizzle AI, which will be acquired as part of the deal for an undisclosed sum.
Sizzle is an educational platform that offers AI-powered tutoring in various subjects, with a particular focus on STEM. The acquisition will integrate Sizzle’s technology into the content that Campus already offers to its user base of 1.7 million students, advancing the company’s vision to provide personalized education.
The deal marks yet another sizable move to bring AI closer to academia – a world which OpenAI seemingly wants to be a part of.
In July, Instructure, which operates Canvas, struck a deal with OpenAI to integrate its models and workflows into its platform, used by 8,000 schools worldwide. The deal enables teachers to create custom chatbots to support instruction.
OpenAI also introduced Study Mode in July, which helps students work through problems step by step, rather than just giving them answers.
While the prospect of personalized education and free tutoring makes AI a draw for the classroom, there are downsides to integrating models into education. For one, these models still face issues with accuracy and privacy, which could present problems in educational contexts.
Educators also run the risk of AI being used for cheating: A report by the Center for Democracy and Technology published last week found that 71% of teachers worry about AI being used for cheating.
TOGETHER WITH JETBRAINS
This AI Agent build for real work
Imagine giving every one of your developers a right-hand man for all of their coding needs – one that explains the logic and reasoning behind each decision, can solve real-world tasks with little to no oversight, and do it all while keeping a holistic view of your entire project.
Meet Junie by JetBrains. This smart AI coding agent takes the time-consuming tasks off your plate so you can focus on the interesting work. Junie understands the structure, logic, and relationships in your code. When you give it a task, it finds the right places to work, the right tools to use, and the best path forward.
MARKETS
SoftBank is building an AI warchest
SoftBank might be deepening its ties with OpenAI. The Japanese investment giant is in talks to borrow $5 billion from global banks for a margin loan secured by its shares in chipmaker Arm, aiming to fund additional investments in OpenAI, Bloomberg reported on Friday.
It marks the latest in a string of major AI investments by SoftBank as the company aims to capitalize on the technology’s boom. Last week, the firm announced its $5.4 billion acquisition of the robotics unit of Swiss engineering firm ABB. It also acquired Ampere Computing, a semiconductor company, in March for $6.5 billion.
But perhaps the biggest beneficiary of SoftBank’s largesse has been OpenAI.
The model maker raised $40 billion in a funding round in late March, the biggest private funding round in history, with SoftBank investing $30 billion as its primary backer.
The companies are also working side by side on Project Stargate, a $500 billion AI data center buildout aimed at bolstering the tech’s development in the U.S.
SoftBank CEO Masayoshi Son has long espoused his vision for Artificial Super Intelligence, or “AI that is ten thousand times more intelligent than human wisdom,” and has targeted a few central areas in driving that charge: AI chips, robots, data centers, and energy, along with continued investment in generative AI.
With OpenAI’s primary mission being its dedication to the development of artificial general intelligence, SoftBank may see the firm as central to its goal.
LINKS
OpenAI’s GPT-5 shows lower levels of political bias than previous models
The Senate passes chip export limits on Nvidia, AMD to China
CoreWeave leaders sold $1 billion in shares after stock hike in August
Elon Musk’s xAI is building world models for gaming
Meta poaches Thinking Machines cofounder Andrew Tulloch
AI videos of dead celebrities are horrifying many of their families
Young people in China are embracing AI therapy
Perplexity pauses new advertising deals to reassess ambitions
Google AI Studio: New update allows users to vibe code using their voice, using AI to strip out filler words and false starts.
JustPaid: AI-powered revenue tool for tracking payments and automating invoice follow-ups.
Intryc: Scores and gives feedback on all support conversations to make sure they’re following the SOP.
Lyra: A meeting platform to rival zoom/meets with built-in recording, transcription and action items.
POLL RESULTS
In 5 years, what do you think will most viral video content be?
Fully AI-generated (29%)
Human-AI collaborations (48%)
Human-only productions (7%)
Impossible to tell (16%)
The Deep View is written by Nat Rubio-Licht, Faris Kojok and The Deep View crew. Please reply with any feedback.
Thanks for reading today’s edition of The Deep View! We’ll see you in the next one.
“The first image is too perfect. Nature has flaws and underwater photography produces difficult lighting conditions; the giveaway was on the tentacles.” “I pulled up actual pictures of octopi and [this image] looked more like them.” “Underwater textures tend to be darker and not that well lit.” |
 | “Suction cups look more natural in their non-similarity to each other.” “They both look fake :D.” “Very hard to tell which is which nowadays.” |
Take The Deep View with you on the go! We’ve got exclusive, in-depth interviews for you on The Deep View: Conversations podcast every Tuesday morning.
If you want to get in front of an audience of 450,000+ developers, business leaders and tech enthusiasts, get in touch with us here.