Bernardo Caldas: "We need to spend more time understanding AI systems in order to control them"
AI News
Искусственный интеллект

Bernardo Caldas: "We need to spend more time understanding AI systems in order to control them"

SIC Notícias10 September 2026 at 06:30

Bernardo Caldas, artificial intelligence specialist, was the guest on the Expresso da Manhã program to discuss the real risks associated with artificial intelligence. The specialist warned that it is necessary to spend more time understanding AI systems in order to adequately control them.

The conversation was led by Paulo Baldaia, who explored with Caldas the implications and dangers of the accelerated development of these technologies. The debate comes in a context of growing worldwide concern about the impact of AI in various areas of society.

The specialist emphasized the importance of deeply understanding artificial intelligence systems before implementing them on a large scale, arguing that only then will it be possible to ensure effective control over these tools.

Related articles

“Estão a construir IA que pode matar-nos a todos em pouco anos”? Responde Bernardo Caldas
AI News

"Are they building AI that can kill us all in a few years"? Bernardo Caldas responds

Bernardo Caldas, artificial intelligence specialist, was interviewed on the Expresso da Manhã program by Paulo Baldaia to discuss the real risks of artificial intelligence. The conversation addressed growing concerns about the development of AI systems and the potential dangers this technology represents for humanity in a short to medium-term horizon, seeking to distinguish between real threats and exaggerated perceptions.

Expresso10/09/26, 06:30
AI News

Warnings from Within

The article addresses the departure of Jacob Coxon, a researcher who worked on the development of ChatGPT and Claude at Anthropic, warning about the risks of advancing toward more powerful AI models without adequate safety or regulation. The article argues that, although the alarm may be debated, it is difficult to ignore these warnings from people who worked in the laboratories building this technology, suggesting that the real danger may not be a science fiction threat, but normalized irresponsibility in the everyday use of AI.

El Periódico Mediterráneo - General10/09/26, 04:00
AI News

Anthropic researcher resigns and warns about dangers of AI development

Jacob Coxon, an Anthropic researcher identified as one of the developers of GPT-4o and GPT-4.5, resigned and warned on social media that those responsible for the exponential development of artificial intelligence believe this technology "could kill us all by the end of this decade," considering that "there is no other human activity that represents this level of threat."

Now10/09/26, 00:58
AI News

AI Redefines the Role of the Programmer, Assesses DKP's Director

The software development market is undergoing an accelerated transformation with the incorporation of artificial intelligence tools into technology teams' work routines. According to industry specialists, rather than reducing the relevance of programming, the trend is for the profession to become more strategic. Professionals are taking on architecture, review, and solution orchestration functions, replacing repetitive tasks that can already be automated. According to a World Economic Forum survey, in 2025, four in ten developers stated that AI had already expanded their career opportunities. Additionally, approximately seven in ten expect their role to change even further throughout 2026. The entity also estimates that 22% of global jobs should undergo structural transformation within five years. This movement is driven mainly by AI advancement at a faster pace than organizations' ability to requalify their workers.

oglobo10/09/26, 00:25