What events in the field of AI took place during the week

Artificial intelligence (AI) continues to rapidly integrate into digital services, from browsers and search engines to navigation and wearable devices. Large technology companies are simultaneously expanding AI functionality and offering users new tools for control, personalization, and data exchange. All the details are in the Izvestia article.

Mozilla will add a switch for all AI functions to Firefox

Mozilla is preparing an update to the Firefox browser, which will feature a full switch to disable AI functions. Users will be able to block both individual tools and all existing and future developments based on generative artificial intelligence.

The release of Firefox 148 is scheduled for February 24. The corresponding section will appear in the desktop version of the browser. For a complete failure, you will need to activate the "Block AI improvements" feature. After that, the system will stop displaying hints and notifications about the capabilities of such tools.

At the same time, users will be able to individually manage functions such as. automatic page translations, alternative text in PDF files, and tab grouping. A chatbot with access to Claude, ChatGPT, Copilot, Gemini and Le Chat Mistral will also appear in the sidebar.

The company emphasized that they rely on the voluntary implementation of technologies. Mozilla CEO Anthony Ensora-DeMeo said that AI should remain an informed choice of the user. This approach distinguishes Firefox from Chrome and Edge, where complete disabling is more difficult.

Google is testing the chat import feature from ChatGPT in Gemini

Google Corporation is testing the import of correspondence from third-party chatbots in Gemini. According to the developers, the feature will allow you to transfer the history of dialogues and user settings. This should simplify the transition between artificial intelligence services.

The function will be called "Import of AI chats". The user will be able to download chat archives through the add files menu. However, the format of the supported data has not yet been disclosed. It is also unclear whether reverse exports from Gemini to other platforms will appear.

Experts note that the initiative raises privacy issues. Transferring correspondence means transferring data between competing companies, but not all users are ready to share their communication history with AI. An additional problem remains the lack of uniform export standards.

Skipping a line: why neural networks were vulnerable to hacking using verses

"Poetic attacks" make AI 62% more likely to break the rules

Yahoo has launched a beta version of the AI Scout search engine

Yahoo has launched a beta version of the Scout search engine based on artificial intelligence. The service generates ready-made structured responses instead of the traditional list of links. The output uses tables, multimedia, and trusted sources. The company claims that this makes information more reliable and understandable.

According to The Verge, this brings the company back to the concept of an "internet travel guide." The engine is based on the Claude model from Anthropic, and the search infrastructure is provided by Microsoft Bing. Scout is currently available in the web version and the Yahoo Search mobile app.

The purpose of the service is to help you find information online and add summaries and smart features to all Yahoo products. Eric Feng, head of Yahoo's research group and leader of the Scout project, noted that the search task has changed. Now it is important not only to find information, but also to cut off clickbait and unreliable AI content.

The technology will be integrated into the company's mail, news, and financial services. This will allow you to generate summaries based on different sources. Monetization is implemented through affiliate links and advertising. CEO Jim Lanzone said that free access remains a priority, but a paid subscription is not excluded in the future.

OpenAI will remove GPT-4o from ChatGPT

OpenAI has announced the removal of GPT-4o from ChatGPT effective February 13, 2026. The GPT-4.1, GPT-4.1 mini and o4-mini will disappear along with the model. However, there are no plans to change the API yet.

The decision is explained by the redistribution of resources. After the temporary return of the model, the developers collected user feedback. It was about warmth of communication and creativity. These parameters were taken into account when creating the GPT-5.1 and GPT-5.2 versions.

According to the company, only 0.1% of users continued to choose GPT-4o on a daily basis, while the main traffic had already switched to GPT-5.2.

In addition, advanced response style settings have been implemented. Users can adjust the tone and degree of emotionality. The goal of OpenAI is to give users more control over the experience of using a chatbot. The company emphasized that personalization has become a key area of development.

A godsend for a spy: how employees of Russian companies give secrets to neural networks

Due to the lack of regulation, employees do not realize that they are transferring sensitive data to the side.

Apple acquired Q.ai for $2 billion

Apple Corporation has acquired an Israeli startup Q.ai for almost $2 billion. This is the second largest purchase in the company's history after the Beats deal. The project team will join Apple divisions.

Developments Q.ai related to machine learning, image and sound processing. The patents describe facial micro-motion recognition technology. It allows you to read silent speech by lip movements.

Experts believe that the solutions can be integrated into AirPods, Apple Watch and Vision Pro. This will allow you to control your devices without voice input. In addition, the technology is being considered for future glasses and headphones with cameras.

Google has added Gemini maps for pedestrians and cyclists

Google is expanding Gemini's integration into navigation services. The voice assistant now works in Google Maps for pedestrians and cyclists. The user can ask questions without being distracted from the route.

Along the way, the AI suggests sights, establishments, and infrastructure. Voice messaging is also available. Currently, the feature is already being rolled out globally on iOS and Android.

Experts note that the technology turns navigation into a personal guide format. However, some experts point to the risk of digital overload, since hiking, first of all, is designed to allow a person to take a break from the screens of gadgets. The final benefit, according to experts, depends on the usage scenario.

All important news is on the Izvestia channel in the MAX messenger.