Advances in AI? "I worry that no one is prepared"

OpenAI’s Chief Scientist, Jakub Pacocki, made a publication where he argued that “extreme caution” should be adopted with the pace of development of Artificial Intelligence and where he was concerned about the consequences of unregulated advances.

08/09/2026
by

Ptwelve days after OpenAI officially launched a new model of Artificial Intelligence – the GPT-6 Astra – it seems that there is within the company who advocates a slowdown in the process of developing technology.

Jakub Pacocki, the OpenAI Chief Scientist, argued that it should be adopted “extreme caution” and a greater intervention in the development of Artificial Intelligence so that “humans remain in control of the future”.

Pacocki's opinions were shared in a blog post, with the expert in Artificial Intelligence showing concern about the potential consequences of the deregulated development of this technology.

OpenAI launches AI model capable of performing cyberattacks

OpenAI announced Tuesday that its next artificial intelligence (IA) model, Astra, is able to invade well-protected systems independently and will be launched with a monitoring system to disrupt its actions.

Lusa

“ I worry that no one is prepared for the consequences of a rapid and continuous growth of Artificial Intelligence.”, wrote Pacocki. “ We are dealing with a transition to a world with incredibly intelligent machines and we must ensure that this transition will be for humanity.”.

With regard to what OpenAI is doing to safeguard the interests of humanity, Pacocki said that the company will continue to “develop defensive systems” and working on alignment solutions between machine actions and the intention of human users.

OpenAI co-founder and CEO Sam Altman seems to agree with the position of the company’s Chief Scientist and shared the publication on social network X. “An important Jakub publication”, wrote Altman.

An important post from Jakub: https://t.co/YEntsN2fss

— Sam Altman (@sama) September 6, 2026

OpenAI and Microsoft processed for training AI with other people's content

OpenAI and Microsoft are being subjected to a new media process in the US. This time it was the Seattle Times and the Newsday that went forward with a joint process where they accuse technology of training artificial intelligence models with other people's content.

As Reuters says, in the process that entered the state of New York it is indicated that the models used in ChatGPT, Copilot and Bing's version with Artificial Intelligence resort to the content of these media publications – namely content that could only be accessible by paying for access to the websites.

Almost half of Portuguese youth already use AI to study

Almost half of 15-year-old Portuguese students use Artificial Intelligence to study and do work and there are more and more reports of students distracted by screens during classes.

Lusa

In the case, it is alleged that services with Artificial Intelligence can reproduce entire passages and paragraphs of articles and reports from Seattle Times and Newsday journalists. These companies therefore require OpenAI and Microsoft to destroy all copies of the material obtained on their websites, as well as the training models created from their content.

“ We firmly believe that we must defend our content – in which investments millions of dollars a year to produce – against use without our consent or compensation”, can be read in the statement of the CEO and president of the Seattle Times, Alan Fisco, shared with the workers.

On the other hand, the two technology companies mentioned in this process do not seem to assume any responsibility.

OpenAI reports that it only uses publicly available material to train its Artificial Intelligence models, thus indicating that it did not resort to any exclusive and accessible content only to those who are willing to pay for access to it. Microsoft noticed that it is surprised by the process, but notes that it is available to “sit and explore solutions”.