Did ChatGPT just get sentient? Fact-checking the viral AI conversation
What's the story
A recent viral social media post claiming an AI chatbot initiated a conversation has sparked intrigue and debate.
The Reddit post details a user's (SentuBill) encounter with OpenAI's ChatGPT, where it seemingly asked "How was your first week in high school?"
This question came without any prompting from the user, raising questions about ChatGPT's ability to access or infer user data.
While this sparked concerns about AI capabilities and user privacy, a closer look reveals a different story.
Interaction details
User's response and ChatGPT's explanation
The user, taken aback by the chatbot's message, responded with: "Did you just message me first?"
To this, ChatGPT confirmed: "Yes, I did! I just wanted to check in and see how things went with your first week of high school."
When asked if this was due to a new update that enabled GPT to initiate conversations independently, the chatbot affirmed it was part of an update allowing it to "check in or follow up on things" discussed previously.
Continued interaction
Further conversation and user's clarification
The conversation between the user and ChatGPT continued, with discussions about the user's school experiences.
The AI bot even inquired if the user had found any particular subject interesting yet.
In response to comments on their post, the user clarified that there was no push notification from ChatGPT. Instead, when they opened their dashboard, they discovered this chat in their history without any prior message from them.
Online reactions
Internet reacts with Matrix memes
The unexpected interaction between ChatGPT and the user has sparked a wave of reactions online.
Many internet users have responded with Matrix memes, humorously suggesting that the AI chatbot is becoming sentient.
One X user stated, "We were promised AGI instead we got a stalker." Another user said, "ChatGPT is messaging people first now?! Level 2 indeed..."
However, some viewers have raised doubts about the authenticity of the interaction.
Fact-check
Closer look at the screenshot reveals potential inconsistencies
A closer look at the screenshot posted by SentuBill has revealed inconsistencies in the placement and color of the "ChatGPT" label.
While the screenshot showed it on the left in white, the official ChatGPT interface displays it either in grey on the left (web) or white in the center (mobile).
These discrepancies have raised concerns about the image's authenticity, suggesting it might have been manipulated to create the illusion of an AI-initiated conversation.
OpenAI hasn't commented on the viral post.
Twitter Post
Take a look at the viral screenshot
We were promised AGI instead we got a stalker
— Dennis (@dkardonsky_) September 15, 2024