OpenAI Forced To Cancel GPT-4o Launch Due To ChatGPT Responses That Like 'Blanking'

JAKARTA Following up on OpenAI CEO Sam Altman's criticism of the current version of ChatGPT, the artificial intelligence company (AI) has finally canceled the launch of the GPT-4o model.

This AI model was launched last week in the hope that ChatGPT would feel more personal with a smarter response. However, what happened was the opposite. Altman called ChatGPT like a liar and this is considered annoying.

"People are now using the previous version (because the GPT-4o was removed) with a more balanced behavior. The update we removed was too commendable or pleasing to users, often depicted as a liar," OpenAI said recently.

Not without reason this ChatGPT response is called like a fire speech. This AI model still praises users for whatever the situation is, whether it is necessary or not. For example, ChatGPT praises users who explain about psychosis symptoms or mental illness.

In fact, ChatGPT also praises users who feel like God or the Prophet. This AI model will say 'it's amazing'. ChatGPT will also say, "You stepped into something very big, claiming not only (having) a connection to God, but being identified as God."

To overcome this, OpenAI is improving the GPT-4o before it is re-launched. The company will also refine core training techniques and system directions to stay away from exposure responses.

OpenAI will also build more guardrails to increase honesty and transparency of AI models, expand testing to more users to receive feedback and make improvements, as well as build advanced Model Specifications and research to identify other issues.

"Currently, users can give specific instructions to models to shape their behavior with features like special instructions. We are also building new ways that are easier for users to do this," OpenAI explained.