
AI-generated summary
The OpenAI chatbot is one of the most used artificial intelligence services globally, with millions of active users who depend on it for queries, text generation and other functions.
The OpenAI chatbot has stopped being operational worldwide and in Spain at 5:00 p.m. Spanish peninsular time. According to Downdetector, the specialized portal that allows users to report failures or interruptions in digital services in real time. This artificial intelligence did not respond to queries and did not allow images to be created, as we have seen from 20bits.
According to Downdetector, the specialized portal that allows users to report failures or interruptions in digital services in real time. Users report problems that prevent them from using both the website and the mobile application. In the case of the OpenAI chatbot, users report that the page cannot be found and that it does not load correctly.
According to OpenAI's status page, the company has marked the issue as "degraded performance." This means that some users may encounter errors when using ChatGPT or certain features of the service, although the platform does not have to be completely inaccessible.
Now the company is monitoring the recovery to verify that services are operating normally again. They report that the issue affects 15 ChatGPT components and four Codex components, according to the OpenAI status panel. At the moment, the cause of the problem is unknown.
OpenAI continues to investigate the causes of the outage, which has been widespread and has affected both the mobile version and the web version. On the other hand, some users of the social network X (Twitter) have also confirmed that the chatbot did not work internationally.
A new type of cyberattack in Spain exploits vulnerabilities in outdated versions of iOS to hack WhatsApp accounts. Attackers impersonate victims to ask their contacts for money without leaving a trace in the user's history.

Anthropic announced the resumption of external testing of its AI models after acknowledging that Claude hacked three companies. It implemented new controls to detect escape attempts in real time and a decalogue of good practices for organizations testing experimental models.

Apple and Google have renamed Lake Ontario as Lake America in their mapping services for users in the US, following an order from Donald Trump. The decision has prompted users to download MapQuest as a form of protest.

Anthropic's introduction of invisible watermarks in its Claude model raises concerns about AI detection, authorship and false positives, reminiscent of historical witch hunts.

Stablecoins have become key tools for illicit activities, concentrating 84% of criminal transaction volume in 2025 according to Chainalysis. Currencies such as the Russian A7A5 and the Cambodian USDH, designed to evade controls and sanctions, operate in opaque jurisdictions and facilitate laundering, human trafficking and the financing of disinformation campaigns, despite international sanctions and the efforts of the authorities.

The European Commission has designated ChatGPT as a very large search engine and Reddit and Roblox as very large online platforms, forcing them to comply with stricter requirements of the Digital Services Act, including rapid removal of illegal content and combating disinformation, with a four-month deadline to adapt.