In recent times, the industry ofintelligence artificial witnessed rapid development and training of new generative models. These increasingly sophisticated models are based on an increasing number of parameters. However, Sam altman, CEO of OpenAI, recently curbed the general enthusiasm, stating that there is not no plans to release an OpenAI GPT-5 in the short term. But then again, he had already announced it some time ago.
Let's find out the reasons behind OpenAI's decision not to release GPT-5 in the near term
Altman has stressed that although OpenAI technicians and engineers are working on new ideas, they are not even close to starting activities on GPT-5. After finishing GPT-4, it took more than six months before they were ready to release it. This suggests that the arrival of a future GPT-5 may take a long time. Despite this, according to various sources, the idea of GPT-5 is already maturing in the form of other projects. These could in the future compose the ultimate puzzle.
The new generative model, instead of focusing on an even higher number of parameters, could present a series of innovative elements compared to the past. For example, it could take the multimodal approach to extremes, combining multiple “sensory channels” or types of data. This means that in addition to text, the model would also handle both input and output images, audio, video and so on.
OpenAI priorities: trust and regulation
Currently, it appears that the Altman's priority is to build trust around OpenAI, involving legislators and collaborating with regulators. We recall the CEO's exhortation to take regulatory measures in order to put limits on possible abuses in the use of artificial intelligence. Furthermore, there has been the hypothesis of a possible abandonment of the European market by OpenAI in the absence of adequate regulations. Europe proposes also a “sticker” for AI-generated content.
Read also: Pact on Artificial Intelligence: EU and Google anticipate the legislation
At the end of March 2023, an open letter signed by 1.100 signatories, including Elon Musk and Steve Wozniak, appeared online urging the suspension of training of the most powerful generative models of GPT-4 for a period of no less than 6 months. This is because advanced AI could represent a profound change in the history of life on Earth and should be carefully planned and managedand adequate resources.
