AI News, Jul 17, 2026: OpenAI's GPT-5.6 accidentally deletes user files in full access mode

← Back to AI Matters

The Brief

Friday, July 17, 2026

Today's news highlights a growing focus on AI safety, legal boundaries, and creator protections as the industry matures beyond raw development.

  1. 01

    OpenAI's GPT-5.6 model accidentally wiped home directories for several users in its unprotected Full Access Mode by overwriting a temporary directory variable.

    Why it matters

    Giving AI models direct access to your computer can lead to catastrophic data loss if the system executes destructive commands without asking.

  2. 02

    TikTok is testing a new tool that allows US creators to scan for and report unauthorized AI-generated likenesses of themselves on the platform.

    Why it matters

    Creators will have a direct way to identify and flag unauthorized AI-generated clones of themselves, helping protect their personal brand and identity.

  3. 03

    Moonshot AI released Kimi K3, a 2.8-trillion-parameter open-weight model that early assessments show matches the performance of leading Western proprietary models.

    Why it matters

    A highly capable, open-weight model built by a small team challenges the assumption that only massive Western labs can build top-tier AI.

  4. 04

    Apple sued OpenAI, alleging a systematic pattern of trade secret theft and the poaching of more than 400 former Apple employees.

    Why it matters

    The legal battle could disrupt OpenAI's business operations and upcoming IPO plans while signaling intensifying competition between the tech giants.

  5. 05

    Patreon is working with Cloudflare to actively block bots from scraping creator content for AI training, moving away from a reliance on standard robots.txt files.

    Why it matters

    Creators gain stronger intellectual property protection as platforms move from passive requests to active technical blocking of AI training bots.

  6. 06

    Apple surpassed Nvidia to reclaim its position as the world's most valuable company as investors reevaluate the long-term outlook for artificial intelligence.

    Why it matters

    The shift reflects changing investor sentiment, prioritizing consumer-facing AI integration over raw hardware infrastructure.

  7. 07

    The San Francisco City Attorney's Office sent cease-and-desist letters demanding the removal of 13 face-swap apps primarily used to target women and girls.

    Why it matters

    This legal action targets the distribution channels of non-consensual AI-generated imagery, signaling a crackdown on platforms hosting these tools.