On July 17, 2026, the official X account of OpenAI unexpectedly published a brief message promoting a version named "GPT-5.6 Sol." Crucially, the post was accompanied by an unofficial domain link, raising significant doubts and concern within the tech community.
Diễn biến chi tiết
Specifically, the post from the @OpenAI account read: "10,000 reasons people love GPT-5.6 Sol" alongside a link redirecting to an unfamiliar website. The link utilized a subdomain "switch-to-codex" under a non-OpenAI address, "openai.chatgpt.site". Immediately after appearing, the post garnered immense attention from developers and artificial intelligence enthusiasts. Many users quickly noticed the anomaly, as OpenAI typically announces major products through formal events or official blog posts on its main domain. Currently, OpenAI has not released any official response regarding the authenticity of the post or whether their account has been compromised by a third party.
Bối cảnh & Nguyên nhân
This is not the first time social media accounts of major tech companies have been targeted by bad actors. In the past, numerous hacking incidents hijacked official X accounts to promote fraudulent cryptocurrency projects or phishing links. The domain structure used in this post exhibits many patterns typical of online phishing campaigns designed to steal user credentials. The mention of an unannounced version like "GPT-5.6 Sol" further supports the hypothesis of a cyber security breach. The community strongly advises users to avoid clicking the shared link to prevent data security risks.
Phân tích kỹ thuật & Công nghệ
Technically, the domain "openai.chatgpt.site" is entirely absent from the official server infrastructure of OpenAI or Microsoft. Typically, OpenAI uses trusted domains such as openai.com or chatgpt.com for all its services and APIs. Utilizing a ".site" top-level domain combined with misleading keywords is a common technique known as "typosquatting" or brand impersonation. Security experts analyze that such websites are often designed to mimic OpenAI's login page to harvest access tokens or banking information. Furthermore, the term "GPT-5.6 Sol" does not align with OpenAI's standard product roadmap, which currently focuses on models like GPT-4o or the advanced o1 series.
Ý kiến chuyên gia & Nhận định
Several independent cybersecurity experts quickly issued warnings on other social media platforms. They noted that the compromise of a verified, gold-badge account with millions of followers like OpenAI's is a major wake-up call for corporate account security. Security researchers emphasized that users must remain highly vigilant against strange links, even when posted by verified, official accounts. Reactions across Reddit and X show that the majority of users recognized this as a potential scam and actively reported the post to prevent widespread reach.
Tác động & Tương lai
This incident once again highlights the vital importance of multi-factor authentication and strict social media access management at large corporations. For Vietnamese users, the key takeaway is to always meticulously verify URL addresses before inputting any personal or account credentials. In the future, platforms like X will need to further tighten security measures to prevent high-profile corporate accounts from being exploited. This security incident could also impact OpenAI's brand reputation if they do not provide a swift explanation and thorough remediation.