OpenAI makes ChatGPT ‘less verbose,’ blurring writer-AI distinction

The enhanced edition of GPT-4 Turbo, which is among the models behind ChatGPT’s conversational capabilities, has been introduced by OpenAI.

With this new enhancement, our AI chat tool can now carry on more conversational interactions that feel more like human exchanges. This improved feature is accessible only to those with ChatGPT Plus, Team, or Enterprise memberships.

The new model from OpenAI, named “gpt-4-turbo-2024-04-09,” comes with improved capabilities in writing, mathematics, logic, and coding, as well as an up-to-date knowledge base.

The latest upgrade was trained on publicly accessible data up to December 2023. The previous version could only draw on material up to April 2023.  OpenAI wrote in a post on X:

“When writing with ChatGPT [with the new GPT-4 Turbo], responses will be more direct, less verbose and use more conversational language.” 

Yet, this modification might make it more difficult for authors to distinguish their work from that of artificial intelligence, which could continue to be a problem.

On April 7th, Paul Graham, co-founder of renowned startup incubator Y Combinator, expressed his belief that an email containing the word “delve” was likely penned using artificial intelligence.

Graham maintained that the term was seldom employed in normal discourse, but this was disputed by numerous Africans, particularly Nigerians, who asserted they frequently utilized it. The discussion has since evolved to highlight how individuals from various regions employ the English language’s vocabulary distinctively.

Graham’s post raises the concern that non-native English speaking writers might face job losses due to their writing being mistakenly identified as being too close in style to content produced by artificial intelligence.

After the introduction of new models on April 10 in OpenAI’s API, such as GPT-4 Turbo with Vision, an update has been released. This update includes advanced capabilities for image comprehension.

Based on information found on its website, OpenAI secretly altered the training data for its GPT-4 Turbo model, effective as of December 2023.

The goal behind the implementation of this update was to address an issue in the model where it failed to finish requested tasks, often referred to by developers as “lack of motivation” or “laziness.”

OpenAI has made it possible for individuals to use its popular AI tool, ChatGPT3.5, without having to create an account as of April 1. This change was implemented to make it more convenient for people to explore the capabilities of artificial intelligence. Yet, users who don’t register won’t be able to save their past conversations.

Read More

2024-04-12 10:54