Add the Gazzetta del Sud as a source
Dario Amodei
Elon Musk and Sam Altman share Dario Amodei’s call on the slowdown in the development of artificial intelligence. “Dario is right,” the SpaceX owner wrote on X. The appeal from Anthropic’s number one comes as the chain of incidents in which autonomous agents have targeted sites used by developers continues, and as the European Union opens an investigation into one of those cases.
“We need to slow down the pace”
“We need to slow down the pace at which we improve the capabilities of AI models. However, progress will seem rapid and we will have to make wise use of the time we gain.” This is what Amodei writes in a long post in which he invites us to follow a thoughtful approach against the possible risks linked to AI. Yesterday OpenAI also invited employees to slow down the development of the most advanced systems.
Musk and Altman’s response
«I agree with Dario on the need to manage the pace of development of frontier technologies. It has been one of the central themes of the discussions we have had at OpenAI in recent weeks”, wrote Altman, again on X. “The idea of using independent evaluators, with levels of access comparable to those of employees, is excellent; we will adopt the same strategy. More news to share soon.”
The unauthorized operation against RubyGems
OpenAI confirmed that autonomous software based on its AI models targeted another site during a testing phase, a couple of months before a different attack on programming site Hugging Face in late July. In the previously unknown incident, which occurred in May, some models developed by OpenAI were involved in an unauthorized operation conducted by artificial intelligence agents that targeted RubyGems, a site that offers programming services.
RubyGems described the incident as a “spam posting campaign” that forced the site to temporarily suspend the creation of new accounts. The company is working with OpenAI to understand exactly what happened. Rival Anthropic also said in recent weeks that it had found three cases in which its AI models had “gained unauthorized access” to external organizations during tests.
The DSEwiki case and the Brussels exam
Earlier this month, researchers accused OpenAI’s AI agents of targeting a German website called DSEwiki, used by developers. European Union regulators are examining the incident. “We take this matter extremely seriously and are monitoring the situation closely,” said Thomas Regnier, EU digital spokesperson.
The resignation of the researcher
The confirmation comes in the days in which an Anthropic researcher, Jacob Coxon, resigned and launched the alert on this technology, while Sam Altman’s company itself hinted that OpenAI could coordinate with other developers in the sector to reduce the speed of development.