OpenAI’s GPT-5.3-Codex Faces California AI Safety Law Scrutiny As Watchdog Alleges High-Risk Violations

Snigdha Gairola

Mon, February 16, 2026 at 9:31 PM GMT+9 3 min read

OpenAI may face significant fines after a watchdog group alleged the company violated California’s new AI safety law with the release of its latest coding model, GPT-5.3-Codex.

High-Risk GPT-5.3-Codex Sparks Safety Concerns

Last week, The Midas Project claimed OpenAI failed to implement legally required safeguards for models classified as high cybersecurity risks, as outlined in its own safety framework.

GPT-5.3-Codex is part of OpenAI’s effort to reclaim its lead in AI-powered coding and, according to company benchmarks, outperforms previous models and competitors in coding tasks.

7/ Despite triggering the “high risk” cybersecurity threshold, OpenAI appears not to have implemented the required safeguards.

Page 29 of their safety report says their measures would not be “adequate for a Safeguards Report”—their framework’s required standard. pic.twitter.com/MukiATYtGG

Don’t Miss:

**The ‘ChatGPT of Marketing' Just Opened a $0.85/Share Round — ****10,000+ Investors Are Already In**

**Own the Characters, Not Just the Content: ****Inside a Fast-Growing Pre-IPO IP Company**

Watchdog Challenges OpenAI’s Compliance Claims

OpenAI disputes the allegations.

An OpenAI spokesperson told Fortune, “GPT-5.3-Codex completed our full testing and governance process, as detailed in the publicly released system card, and did not demonstrate long-range autonomy capabilities based on proxy evaluations and confirmed by internal expert judgments, including from our Safety Advisory Group.”

However, Midas Project founder Tyler Johnston criticized the release.

This is _“_especially embarrassing given how low the floor SB 53 sets is: basically just adopt a voluntary safety plan of your choice and communicate honestly about it, changing it as needed, but not violating or lying about it," he said.

Safety researchers like Nathan Calvin of Encode also questioned OpenAI’s defense, noting, “From reading the relevant docs … it doesn’t look ambiguous to me.”

Trending: Blue-chip art has historically outpaced the S&P 500 since 1995, and fractional investing is now opening this institutional asset class to everyday investors.

OpenAI Growth Surges Amid Anthropic Rivalry

On Monday, OpenAI’s momentum increased as CEO Sam Altman reassured employees and investors despite mounting competition from Anthropic’s upgraded coding tools.

Altman said in an internal Slack message that ChatGPT had returned to more than 10% monthly growth and that an updated Chat model was set for release that week.

Story Continues  

He also noted that Codex usage rose about 50% after the launch of GPT-5.3-Codex and a standalone Mac app.

OpenAI pushed back against Anthropic’s Super Bowl ads, criticizing the idea of advertising in ChatGPT.

Altman called the ads “deceptive,” though a person familiar with the matter said the company planned to test clearly labeled ads placed at the bottom of responses without influencing answers.

**Read Next: **

**Wall Street's $12B Real Estate Manager Is Opening Its Doors to Individual Investors — ****Without the Crowdfunding Middlemen**

**This ETF issuer isn't chasing the index — ****it's building tools for income, leverage, and conviction**

Photo courtesy: Shutterstock

UNLOCKED: 5 NEW TRADES EVERY WEEK. Click now to get top trade ideas daily, plus unlimited access to cutting-edge tools and strategies to gain an edge in the markets.

Get the latest stock analysis from Benzinga:

APPLE (AAPL): Free Stock Analysis Report

TESLA (TSLA): Free Stock Analysis Report

This article OpenAI’s GPT-5.3-Codex Faces California AI Safety Law Scrutiny As Watchdog Alleges High-Risk Violations originally appeared on Benzinga.com

Terms and Privacy Policy

Privacy Dashboard

More Info