OpenAI claims to have fixed ChatGPT's biggest problem

Since its debut in late 2022, ChatGPT has sparked conversations about the fine line between AI-generated and human-written content. Users quickly identified several markers unique to AI, with the frequent use of em dashes being the most evident. Despite attempts to instruct ChatGPT to avoid these punctuation marks, the issue persisted. In a recent update, OpenAI's CEO, Sam Altman, announced that the company has addressed this concern. Users can now confidently request that ChatGPT refrain from using em dashes, and the AI will comply. "Small-but-happy win: If you tell ChatGPT not to use em dashes in your custom instructions, it finally does what it's supposed to do!" Altman shared on social media. Testing the update, ChatGPT successfully responded without using any dashes, affirming its new capability. However, some users responded to Altman’s announcement by pointing out that the issue still exists for them, indicating that not every instance has been resolved. The challenges of AI hallucinations were also highlighted when users noted that ChatGPT incorrectly identified Joe Biden as the current President of the United States, underscoring ongoing accuracy issues in AI responses. In addition to fixing its em dash dilemma, OpenAI unveiled a new feature for ChatGPT called Group Chats. This addition aims to introduce a social element similar to WhatsApp Groups, allowing users to interact while including ChatGPT as a participant. Up to 20 people can join these conversations through shared links, although this feature is currently in its pilot phase and is being rolled out selectively. This is not OpenAI's first foray into new territories; just weeks ago, the company launched Sora, an AI-driven social media platform for creating and sharing short videos akin to TikTok. Furthermore, last month, it introduced Atlas, an AI browser tailored for Mac users, positioning itself in competition with existing browsers like Google Chrome, Brave, and Comet.