ChatGPT-4o: Is your data safe?
In May of 2024, OpenAI — the company behind numerous AI models such as ChatGPT and DALL-E — announced their newest GPT model: GPT-4o. Recent corporate decisions make us wonder: is ChatGPT-4o safe?
What is ChatGPT-4o?
ChatGPT-4o is OpenAI’s newest GPT model. Expanding upon previous iterations, ChatGPT-4o — with the ‘o’ standing for ‘omni’ — provides a surplus of new features to GPTs. With an output speed of 86.9 tokens/second, GPT-4o is substantially faster than its predecessors. GPT-4o also provides a more friendly user interface, especially on desktop browsers. Additionally, similar to Pi-AI (2022), you are now able to speak to ChatGPT with your voice, and it will respond in a human-like voice and tone, at surprisingly fast rates.
These features are great, however, the commonly held question among users has still not been answered: is my data safe?
Non-profit to for-profit
OpenAI is a non-profit organization. ChatGPT-4 was always a paid subscription. While making GPT-4o free for all users may show signs of OpenAI emphasizing its commitment to AI inclusivity and pioneering the front of AI research, the opposite may be true as well: by providing GPT-4o for free but limiting its use, this may be a strategy to entice more users to purchase the full capabilities, after realizing the benefits over GPT-3.5. The freemium model, that OpenAI is now employing, is a marketing strategy used by many for-profit organizations but is rarely seen in non-profits.
OpenAI’s shift to a for-profit organization is further confirmed by its partnership with Microsoft. In addition, OpenAI’s CEO, Sam Altman, later announced that OpenAI will be making a transition to a for-profit company.
The transition to a for-profit organization seems to show that OpenAI has deprioritized creating secure AI systems and now focuses on solely profit.
Is your data safe?
Short answer — no.
Long answer — it’s no secret that companies extract user data. When announcing AI features, companies must be extremely careful and address all concerns related to data privacy. In June of 2024, Apple announced Apple Intelligence: a brand new AI integration into Apple devices. Apple made sure to address privacy concerns and guaranteed that all Apple Intelligence tasks were completed on-device.
However, OpenAI has not addressed privacy concerns regarding its latest model ChatGPT-4o. Even worse, ChatGPT now has a desktop app, which can leak sensitive business data and financial records. Recently, former co-chair Elon Musk has expressed his concerns regarding the model’s security measures stating that “if Apple integrates OpenAI at the OS level, then Apple devices will be banned at my companies. That is an unacceptable security violation”.
In light of this, it is best to never share personal information — such as bank information, passwords, or addresses — on the platform. ChatGPT is still known as an antithesis of Explainable AI, meaning its responses are not explainable and could be breaching privacy, and while ChatGPT is still safe to use casually, never share personal information.
