Sam Altman defends the availability of artificial intelligence to the public and Amazon launches a decision-making model

Sam Altman, CEO of OpenAI, said that the benefits of artificial intelligence justify accepting some of the risks associated with it, considering that this technology should remain widely available to the public.

“We believe that the world has to accept that some negative things will happen in exchange for the benefits that this technology provides, while giving people the ability to use it,” Altman said in an interview with Politico’s technology publication Decoded.

He added that there is a fundamental difference in vision between “OpenAI” and its competing company, “Anthropic,” regarding the regulation of artificial intelligence, saying: “I think the gap between us is large.”

He continued: “I understand the point of view of those who say that this technology will become so powerful and dangerous that one laboratory in San Francisco should monopolize it and ensure that no harm occurs, and then decide how to distribute its benefits.”

He added: “But I do not agree with this opinion, as this proposal involves a completely unacceptable trade-off,” stressing that it conflicts with the approach of easing regulatory procedures that OpenAI supports.

He said: “I will not accept a trade-off based on saying: We will ensure that there are no major hacks, misuse of this technology, fraud, or other negative things that may occur.”

He added: “I believe that people will use this technology for positive things that are many times greater than the harm that may result from it.”

Altman's statements come at a time when controversy is escalating within the artificial intelligence sector regarding the pace of development of these technologies and the risks that may arise from increasingly capable systems.

Anthropic CEO Dario Amodei in September published an article calling for the sector to slow the pace of development, a position Altman has publicly said he supports.

Anthropic researcher Jacob Cookson resigned in September, saying those working on developing artificial intelligence believe it “could kill us all by the end of the decade.”

Amazon Web Services (AWS), a subsidiary of the American e-commerce and technology giant Amazon.com, has launched an open-source artificial intelligence model for decision-making inspired by TypeSafe’s “GIF” model, as artificial intelligence developers increasingly seek intelligent systems that are more suited to computer automation than highly sophisticated large language models.

The TechCrunch website, which specializes in technology topics, pointed out that the artificial intelligence model for decision-making, “Strands Decider 2B,” which was launched by “Amazon” coincided with a similar product revealed by the artificial intelligence application company “OpenAI.”

This model is a quick and low-cost way to compare pre-determined options, while providing an indicator of confidence in the proposed decision. The Strands Decider 2B model is completely open source and is currently available, allowing it to be used with applications produced by companies other than Amazon. At the same time, it is small in size so that users can run it directly on their devices without the need for external servers.

The TechCrunch website said that Mark Brooker, an engineer at Amazon, created this model after learning about the “GIF” model and trying to build his own version of this type of model. The project - which began as a personal initiative - achieved remarkable success, briefly topping the Jevbench ranking of similarly sized models, which prompted Amazon engineers to improve it and launch it as a product within Strand Labs, a unit of the company that develops new tools and protocols for deploying artificial intelligence agents. Brooker points out that the need for a tool like this arose during discussions with Amazon Web Services customers, whose businesses do not always need to use artificial intelligence agents. Artificial tools that require the high capacity or high cost associated with large language models, with all their capabilities.

A former OpenAI safety employee who recently resigned criticized the company's approach to AI safety, saying that an accelerating culture focused on rapid development increases the risk of failures.

In an article published in The Atlantic magazine today, Saturday, entitled “I resigned from OpenAI because its culture was flawed,” David Robinson said that artificial intelligence companies, including OpenAI, are not being cautious enough and should pay more attention to experience and research in the field of safety before developing systems with greater capabilities.

“The era of trial and error is over,” Robinson said, stressing that advanced artificial intelligence systems require safeguards closer to those used in sectors such as nuclear energy and aviation.

Robinson added that OpenAI relies largely on what it calls “iterative deployment,” whereby it releases systems and enhances protections when problems arise.

These comments add to the ongoing debate within the artificial intelligence sector about whether companies are moving too quickly to develop systems with increased capabilities. OpenAI and its competitor, Anthropic, have faced intense scrutiny following incidents in which safety controls failed or experimental systems behaved unexpectedly.

As the company rushes from one launch to the next, it fails to achieve the level of care that “I believe is necessary,” wrote Robinson, who said he spent three-and-a-half years at OpenAI, helped craft the company’s preparedness framework and oversaw safety reporting on 12 pilot launches.

An OpenAI spokesperson said in a statement: “We ensure that our models do not become more capable than we can safely manage or secure, and we pause training or postpone the launch of models when we need to slow down the pace of work.”

Robinson warned that AI capabilities are advancing faster than researchers understand the concept of “alignment,” a field focused on ensuring that AI systems perform in accordance with humans’ goals and values.