ChatGPT, the famous conversational agent developed by the company OpenAI, has just received a major update called GPT-4 Turbo. This new version promises more natural and fluid conversations, as well as easier access for users. Here’s what you need to remember about the latest developments in this very popular artificial intelligence tool.
ChatGPT: an improved language model for more direct exchanges
The latest iteration of ChatGPT, based on the GPT-4 Turbo language model in its version dated April 9, 2024, brings its share of improvements. According to OpenAI, the generated responses will now be “more direct, less wordy and will use more conversational language“.
In addition to better command of the language, GPT-4 Turbo has an extensive knowledge base, based on public data up to December 2023, compared to April 2023 for the previous version. Progress is also to be noted in the areas ofwritingof the mathematicsof logical reasoning and some programming.
Conversational AI accessible without an account
Another important new feature: it is no longer necessary to create an account to use the chatbot. Since April 1, OpenAI has removed this obligation in order to “facilitate the experience of the potential of AI” for as many people as possible. However, unregistered users will not be able to save their interaction history.
Premium subscribers to ChatGPT Plus plans, Team or Enterprise benefit as a priority from these developments. They can already chat with the new version of GPT-4 Turbo and benefit from its extended capabilities.
Towards a blurrier border between human and AI-generated texts?
While these improvements are impressive from a technological perspective, they also raise concerns. With increasingly natural responses, it is becoming increasingly difficult to distinguish text written by a human from content generated by artificial intelligence.
This problem was recently highlighted by Paul Grahamco-founder of the startup accelerator Y Combinator. In a tweet, he claimed that an email containing the word “delve” (to deepen in French) had necessarily been written with the help of an AI, this term not being, according to him, common in the conversations of all days.
This statement sparked strong reactions, particularly from African Internet users, mainly Nigerians, for whom this word is part of the usual vocabulary.
Beyond the question of linguistic diversity, this episode illustrates the risks of discrimination that the democratization of conversational AI could generate. Writers who speak perfect English but have lexical specificities linked to their origin could thus see their work wrongly assimilated to artificial production.
New features in development at OpenAI
The Californian firm does not intend to stop there and is already working on the next developments of its star conversational agent. On April 10, it launched new models in its APInotably a version of GPT-4 Turbo integrating image understanding functionalities.
The OpenAI teams are also working to reduce what they call “laziness” of the model, namely his propensity to not complete the tasks requested. By refining the training dataset, they hope to obtain more complete and relevant results.
As ChatGPT continues to gain in power and accessibility, the question of its impact on our societies arises with increasing urgency. While this technology opens up tremendous prospects, it also carries risks that must be anticipated and managed. A major challenge for OpenAI and the entire conversational AI ecosystem.
My name is Ethan, I’m 30 years old, and I’m the founder of this magazine. After studying journalism, I quickly decided to go freelance to be able to write as I wanted and talk about my true passions. I aim to create an information hub where French and international news is treated seriously and with insight. My goal is to enlighten my readers’ daily lives and encourage a deeper understanding of current issues.
