Article illustration 1

OpenAI has launched a major overhaul of ChatGPT's personalization settings, introducing granular controls for the AI's personality and communication style—but the update is facing significant user backlash. The new hub, previewed by CEO Sam Altman on X, allows users to select from predefined personas like "Cynic," "Robot," or "Nerd," alongside custom instructions for tailoring responses. Memory toggles and fields for specifying nicknames, occupations, and personal preferences aim to create a more individualized experience.

Personality as a Product

This push for personalization arrives after GPT-5's lukewarm August debut, where users criticized its slower performance and perceived emotional flatness compared to GPT-4o. Altman acknowledged the feedback, stating: "We understand there isn't one model that works for everyone" and emphasizing investments in "steerability research." The move reflects a broader industry trend where AI personality engineering has become critical for adoption—whether for Grok's edgy tone or Claude's cautious professionalism.

The 4o Factor

Despite the new controls, many users argue the update sidesteps core issues. On social media, demands to permanently reinstate GPT-4o for all tiers have intensified. One commenter noted: "No combination of toggles replicates the organic vibe of 4o." Others contend OpenAI misunderstands the critique—valuing intuitive adaptability over manual customization. The discontent mirrors recent outcry over plans to retire ChatGPT's Standard Voice Mode, which OpenAI later reversed.

The Uncanny Valley of AI Interaction

This tension underscores a fundamental challenge in generative AI: balancing accuracy with relatable communication. GPT-5's initial rigidity—and the new hub's prescribed personalities—highlight how difficult it is to simulate authentic interaction. As developers, we're reminded that UX isn't just about features; it's about perceived empathy and responsiveness. For teams building atop OpenAI's API, these updates signal both opportunity (customizable agents) and caution (user expectations for "human-like" fluidity).

The personalization hub represents OpenAI's bet that configurability can bridge diverse user preferences. Yet if the backlash persists, it may need to reconsider whether steering wheels compensate for an engine users didn't ask for—or if sometimes, the best interface is simply one that disappears.

Source: ZDNET (Webb Wright, September 17, 2025)