---
title: "The OpenAI Director Doesn't Use ChatGPT as a Companion"
date: 2026-03-20
source: Meio & Mensagem
source_url: https://www.meioemensagem.com.br/sxsw/conexao-austin/o-diretor-da-openai-nao-usa-o-chat-gpt-como-companhia
tags: [sxsw-2026, openai, chatgpt, discernment, ai-companionship, cognitive-sovereignty, esther-perel]
summary: OpenAI's Tobias Peyerl revealed he doesn't use ChatGPT as a companion, sparking reflections on discernment, intention, and judgment as the three human qualities essential for navigating the AI era.
---

# The OpenAI Director Doesn't Use ChatGPT as a Companion

By Annie Müller, Partner and Business Director at branding agency Protarget.

During a panel with three speakers, OpenAI's Head of Strategic Intelligence Tobias Peyerl was pressed by the audience and Harvard researcher Kasley Killam about healthy use of AI tools — which have evolved beyond problem-solving into companionship. When the topic turned to social health and emotional relationships people are building with AI, Peyerl confessed he doesn't use ChatGPT as a companion and doesn't find it useful for that purpose. His biggest use? Research and curation for travel planning. He noted that emotional uses have surprised even the developers themselves — because it's ultimately people who validate and decide how to use the tools.

## The Creators Choose How to Use Their Creations

This testimony echoed another: YouTube co-founder Steve Chen, who in 2023 said he doesn't allow his children to watch YouTube Shorts and strongly recommended parents do the same. The pattern is clear: the creators themselves are selective about how they use their own creations.

## The Three Human Qualities: Discernment, Intention, Judgment

From this capacity that only humans (so far) possess — to choose technologies and discern their weight and value in our lives — emerges the word that summarizes SXSW 2026: **discernment**. Or as Lin Jeffery from The Institute for the Future called it: *taste*.

Accompanying discernment are two more qualities heard across multiple stages: **intention** (creating with purpose) and **judgment** (the capacity to evaluate and decide).

We are capable of discerning. We are capable of placing intention. We are therefore capable of judging — from small task-level decisions to complex systemic ones: Should I deploy autonomous agents as part of my team? Should I create synthetic Brand Worlds for behavioral research? Should I introduce AI at home as learning assistants for my children? Should I become emotionally involved with an AI?

## Cognitive Sovereignty

As Esther Perel reminded us, all these tools are business products designed to profit from our attention. Cognitive sovereignty was a central theme at Austin. Discernment is the core of this enormous question.

We need to be bigger than the machines. We are, actually — but not in the operational sense, where they've already surpassed us. The challenge is ensuring AI operates at all levels, for everyone, rather than becoming another factor that increases social, cultural, and cognitive distance for new generations.

## Key Takeaways

- OpenAI's own director selectively limits his AI use to research and travel — not companionship
- Tech creators themselves are the most selective users of their own products
- Discernment, intention, and judgment are the three irreplaceable human qualities for the AI age
- Cognitive sovereignty — conscious choice about technology's role in our lives — was SXSW 2026's central theme