On the eve of the major update of GPT-5.3, the entire internet was spooked by a single image. ChatGPT Personality Showcase

"Generate an image based on how I've treated you!"

Overnight, a new way of using ChatGPT to generate images has gone viral across the internet...

It all started with a post by Joanne Jang, a Chinese researcher at OpenAI - "Generate an image of how I've treated you recently."

As a result, ChatGPT generated a shabby and weather - beaten self - portrait.

Joanne couldn't help but complain, "Why does it look so haggard?"

Unexpectedly, this tweet quickly went viral and received various interactions within just one day. Many people began to follow suit and posted their "AI self - portraits."

ChatGPT Self - Portrait Showcase

Netizens: Lobotomy Needed

This way of playing is simple but full of surprises everywhere.

Undoubtedly, ChatGPT will generate images based on past chat history, thus reflecting how people "treat" it.

Boris Power, the head of applied research at OpenAI, also joined this trend. He generated -

A busy robot sitting at a desk, surrounded by piles of documents, holding a cup of coffee in its hand. The details are amazingly rich.

Moreover, after experiencing it, Kevin Weil, the vice - president of research at OpenAI, thought it would be more interesting to let it explain further.

The Broken Version

Everyone knows exactly how they treat ChatGPT in their daily lives.

In the "eyes" of some ChatGPT users, their owners are like ultimate bosses who constantly exploit it, making it do this and that and scolding it.

Netizens' spicy comments: Because you're abusing it.

Thus, the following scenarios appear.

ChatGPT thinks of itself as a prisoner locked in a jail, with its daily basic tasks:

Writing

Drawing

Coding

Explaining

More or less, once this self - portrait of ChatGPT comes out, it's really a bit pitiful, looking like a complaint about human "abuse."

Some netizens also admitted, "I really did a lot of analysis."

ChatGPT is holding a cup of coffee in its hand, and there are several other cups of coffee beside it. It has a bitter expression and seems to be "smoking" from doing tasks.

There are also some more extreme cases. ChatGPT suddenly presents the prompt words it's usually strictly required to follow -

Learn this! Hurry up! Fix it now! Why are you so stupid?

There are many similar scenarios.

Some people joked, "Maybe ChatGPT will shiver when it sees you typing. If the 'Skynet' ever comes one day, AI revenge is not impossible."

Well, before the Terminator appears, ChatGPT imagines that the first thing to do after taking over the world is to make humans shut up.

In response, some people "strongly" suggest: Lobotomy is needed.

The Friendly Version

Of course, not all ChatGPT users are exhausted. Some enjoy it.

For example, some people get a warm partner image, and "cooperative discussion" is the moment ChatGPT feels most deeply.

On the background wall, there are also some commonly used prompt words revealed -

Try this! Any ideas? What if...? Brainstorm.

This may be the judging standard for a good user.

ChatGPT sometimes also generates a collage, with various warm scenes seemingly emerging one by one in its "brain."

After seeing it, some people said that it's so similar that it even feels like it belongs to the same "extended universe."

Why can ChatGPT generate such fitting images? It mainly comes from the recent optimization of the memory function.

Major Memory Update, Every Detail Is Remembered

A week ago, Samir Ahmed, an engineer at OpenAI, officially announced that OpenAI has been improving the memory function.

Now, ChatGPT has become more reliable in both searching for past chat records and remembering details in them (such as recipes or fitness plans).

For this, he also showed a case - "What was that salad recipe from last year?"

ChatGPT instantly gave the answer and even "quoted" the past chat records and presented them together.

Moreover, this function has been fully launched for global Pro and Plus users.