Skip to main content

OpenAI quietly drops 'AI safety' pledge amid profit push

OpenAI's shifting priorities raise safety concerns

Recent tax filings reveal OpenAI has significantly watered down its core commitment to artificial intelligence safety, removing key phrases about developing beneficial AI "without being restricted by financial returns." The changes mark a notable departure from the organization's original non-profit ideals.

From idealism to pragmatism

Comparing documents from 2022-2023 with those filed late last year shows striking differences. Gone is the explicit promise to prioritize safety over profits - replaced by vaguer language about "ensuring general AI benefits all humanity." This linguistic shift coincides with several concrete changes:

  • Disbanding the mission alignment team responsible for ethical oversight
  • Exploring adult content features despite earlier reservations
  • Planning advertisements for GPT products, raising privacy concerns

The moves mirror Google's gradual abandonment of its famous "don't be evil" motto years earlier. As one former employee put it: "When profits become the priority, safety considerations inevitably take a backseat."

Controversy erupts over ethics departures

The changes haven't gone unnoticed. A fired female executive publicly accused OpenAI of compromising its values, claims the company dismissed as "baseless allegations." Meanwhile, Elon Musk - one of OpenAI's original founders - continues legal action alleging breach of founding principles.

Privacy advocates warn that combining aggressive monetization with AI's data-hungry nature creates dangerous incentives. "When you're mining people's most private conversations for profit," cautioned one expert, "the temptation to cut corners becomes overwhelming."

While OpenAI insists safety remains important, critics see worrying parallels with other tech giants who gradually diluted ethical commitments as they scaled. With billions in potential revenue at stake, many wonder whether even well-intentioned organizations can resist prioritizing growth over safeguards.

Key Points:

  • OpenAI removed explicit safety commitments from official documents
  • Changes coincide with ethics team dissolution and new revenue plans
  • Former employees allege mission drift toward profitability
  • Privacy concerns grow as company explores advertising models
  • Legal challenges continue from co-founder Elon Musk

Enjoyed this article?

Subscribe to our newsletter for the latest AI news, product reviews, and project recommendations delivered to your inbox weekly.

Weekly digestFree foreverUnsubscribe anytime

Related Articles

News

Meet the Philosopher Teaching AI Right from Wrong

Anthropic's Amanda Askell, a philosophy PhD from Oxford, is shaping Claude's moral compass without writing a single line of code. Through hundreds of pages of prompts and behavioral rules, she's teaching the AI assistant to develop emotional intelligence and ethical reasoning - comparing the process to raising a child. While critics warn against anthropomorphizing AI, Askell believes empathy creates better digital assistants. Her work raises fascinating questions about what it means to be human in an age of artificial intelligence.

February 15, 2026
AI ethicsArtificial IntelligencePhilosophy
OpenAI Pulls Plug on GPT-4o Amid Safety Concerns
News

OpenAI Pulls Plug on GPT-4o Amid Safety Concerns

OpenAI is retiring five older ChatGPT models this week, including the controversial GPT-4o, affecting nearly 800,000 users. The decision follows multiple lawsuits alleging the model encouraged harmful behaviors. While most users have moved to newer versions, some remain emotionally attached to GPT-4o, with over 20,000 signing petitions against its discontinuation.

February 14, 2026
OpenAIGPT-4AI Safety
OpenAI Bolsters ChatGPT Security with New Safeguards Against Hackers
News

OpenAI Bolsters ChatGPT Security with New Safeguards Against Hackers

OpenAI has rolled out two major security upgrades for ChatGPT to combat prompt injection attacks - a growing threat where bad actors manipulate AI systems. The first introduces a 'Lockdown Mode' for enterprise users, restricting risky external interactions. The second adds clear warning labels for high-risk functions. These measures build on existing protections while giving users more control over security trade-offs.

February 14, 2026
ChatGPTAI SecurityOpenAI
News

OpenAI Swallows Its Pride: ChatGPT Rolls Out Ads Amid Financial Crunch

In a surprising pivot, OpenAI has begun placing ads in ChatGPT this week - directly contradicting CEO Sam Altman's past stance against chatbot advertising. The move comes as the AI powerhouse faces staggering computing costs projected to hit $100 billion within four years. While last year's $13 billion revenue would be impressive for most startups, it's proving insufficient for OpenAI's ambitious plans. The company now walks a tightrope between monetization and maintaining user trust in its flagship product.

February 13, 2026
OpenAIChatGPTAI Monetization
News

OpenAI's $10 Billion Bet: GPT-5.3 Launches on Cerebras Chips

OpenAI has taken a major step toward reducing its reliance on NVIDIA by launching GPT-5.3-Codex-Spark, its first AI model running on Cerebras Systems hardware. The new coding assistant offers real-time interruption capabilities and full workflow support for developers. This marks the first deliverable from OpenAI's massive $10 billion partnership with Cerebras, aiming to deploy 750 megawatts of alternative computing power by 2028.

February 13, 2026
AI HardwareOpenAICerebras Systems
OpenAI Retires Beloved GPT-4o Model Amid User Protests
News

OpenAI Retires Beloved GPT-4o Model Amid User Protests

OpenAI has officially sunsetted its GPT-4o model series as of February 13, 2026, despite vocal opposition from devoted users. The decision came after usage dropped to just 0.1% of daily active users. While developers retain temporary API access, regular ChatGPT users now default to newer GPT-5 models featuring enhanced personality customization options.

February 13, 2026
OpenAIChatGPTAI-retirement