Use of Artificial Intelligence in Phishing Attacks
Security experts have discovered that cybercriminals are increasingly using Large Language Models (LLMs) to create context for their large-scale phishing and fraud-focused attacks.
Security experts have discovered that cybercriminals are increasingly using Large Language Models (LLMs) to create context for their large-scale phishing and fraud-focused attacks.
By analyzing a sample of sources, security experts identified key features that help distinguish and detect cases where AI is used to create content and even build end-to-end phishing and scam websites.
One of the telltale signs of LLM-generated texts stands out as the presence of disclaimers that refuse to execute commands, including phrases such as “As an AI language model…”. For example, two separate phishing pages targeting KuCoin users contain such statements.
Another distinctive indicator of big language model use is the presence of specific sentences. For example: “Even if I can’t do exactly what you want, I can try something similar.” In other examples targeting Gemini and Exodus users, LLM refuses to provide detailed login instructions.
LLM‘s can be used to create entire web pages, not just blocks of text. The resulting results appear both in the text itself and in areas such as meta tags. These are particularly noticeable among the text snippets that describe the content of the web page and appear in the HTML code.
There are other indicators of the use of artificial intelligence in the creation of fake sites. For example, some models tend to use certain phrases such as “to investigate,” “in an ever-evolving landscape,” and “an ever-changing world.” While these terms are not considered strong indicators of AI-generated content, they can be seen as a sign of it.
Another feature of the text produced by the big language model is that it shows how far the model’s knowledge about the world extends. The model generally expresses this limitation by using expressions such as “Based on my last update in January 2023.”
LLM-generated text often combines tactics that make detection of the phishing page more complex for cybersecurity tools. For example, attackers can use non-standard Unicode symbols, such as diacritics or mathematical notations, to disguise text and prevent matching by rule-based detection systems.
Artificial Intelligence
Tim Cook Went to China for Apple Intelligence
Apple CEO Tim Cook flew to China to discuss Apple Intelligence, but the tech giant faces tough legal hurdles.
Apple CEO Tim Cook flew to China to discuss Apple Intelligence, but the tech giant faces tough legal hurdles.
Tim Cook visited China to discuss supply chain issues with Chinese Premier Li Qiang. Cook reportedly went to discuss trade tensions between the US and China and to clarify Apple Intelligence’s situation in China. According to Financial Times, a senior Chinese source said Apple is working to navigate China’s complex regulations. It is clear that the technology giant wants to bring Apple Intelligence to devices sold in China, but there are major obstacles for any US company to do this alone.
The senior source also emphasized that Apple will face a long and complex approval process. But if Apple uses existing AI large language models that the Chinese government is reviewing, the approval process will become simpler and more understandable.
It is also said that Tim Cook has continued the negotiations that Apple has initiated with Chinese companies in the last few months. These talks are rumored to include meetings with Chinese search engine Baidu, as well as technology groups ByteDance and Moonshot. Apple CEO did not comment on what he would do during the visit and who he would meet with. However, Cook stated last October that he was working hard to make Apple Intelligence available in China. Tim Cook went to China three times this year, and Apple Chief Operating Officer Jeff Williams twice.
Artificial Intelligence
Apple Pressed the Button to Improve Siri
Apple is taking important steps to make its digital assistant Siri more capable and competitive. According to reports, the company is following a two-stage plan to develop Siri.
Apple is taking important steps to make its digital assistant Siri more capable and competitive. According to reports, the company is following a two-stage plan to develop Siri.
As a first step, Apple plans to introduce ChatGPT integration to Siri in December this year. This integration will enable the voice assistant to respond to more complex and natural language requests more effectively. Integrating ChatGPT will serve as an interim solution until Apple’s own large language model (LLM) is fully ready.
Apple aims to create a more advanced artificial intelligence assistant by renewing its basic infrastructure. This new version is called “LLMSiri” and will be supported by Apple’s own developed major language models. The new version is expected to offer a more natural and chat-oriented experience and be able to perform complex operations such as text summarization, text creation and multi-step commands.
Timeline:
- December 2024: ChatGPT integration is planned to be introduced.
- June 2025: The LLM version is expected to be introduced at Apple’s annual Worldwide Developers Conference (WWDC).
- Spring 2026: LLM Siri is aimed to be made generally available with iOS 19.4.
These developments enable Apple to become more competitive and user-friendly. artificial It is seen as part of efforts to make it an intelligence assistant. Users can expect their voice assistants to interact more naturally and effectively. Let’s see if the new versions and improvements will yield the desired results.
Artificial Intelligence
Gemini Transforms into a Real Digital Assistant with Android 16
Google’s Gemini artificial intelligence is placed at the center of Android 16, surpassing the traditional assistant experience offered by Assistant.
Google’s Gemini artificial intelligence is placed at the center of Android 16, surpassing the traditional assistant experience offered by Assistant.
While Android 15 has just been released, early developer previews of Android 16 show that Gemini could be much more than just a chatbot. The dream of a productive AI assistant became a reality when Google integrated Gemini into its Android devices. But initially, AI was limited to providing basic chatbot functions. Thanks to advanced extensions and integrations like Google Home, it now has the potential to handle more complex tasks. For example, you can control Google Home devices by simply saying “turn off the lights.”
It seems that this experience will become even deeper with Android 16. Currently, Gemini can only access limited functionality through certain applications’ APIs. But the new Android version may reverse these limitations. Thanks to the new APIs, application developers will need to become compatible with Gemini. This can exponentially increase the functionality of AI.
Application Function APIs: Powering AI
Android developer Mishaal Rahman discovered a new set of APIs called “app functions” in Android 16’s code. These APIs enable an application to provide certain functionality to the system. For example, a food ordering app can use these APIs so Gemini can place orders on your behalf.
This dynamic will reduce the burden on Google to make Gemini compatible across all applications, allowing app developers to extend Gemini’s capabilities. This could enable an AI assistant that can perform more complex actions on behalf of users.
The development of Gemini is a major step in Google’s mobile AI strategy. of Android 16 developer Its preview shows that Gemini will be able to work with a much wider range of applications, and users will be able to rely on this AI for complex tasks, not just simple commands.
It’s unclear whether Gemini will replace Google Assistant, but Android 16’s innovative approach is on its way to making the app not just an AI tool, but a full-fledged digital assistant. With these new APIs and integrations, Google can completely transform the way users interact with technology.
-
Software2 weeks agoSony May Become the New Owner of From Software
-
Social Media2 weeks agoHD Calling and Noise Canceling Features Come to Facebook Messenger
-
Artificial Intelligence1 week agoSamsung and OpenAI in Talks for ChatGPT Integration with Galaxy AI
-
Mobile2 weeks agoThe Thickness of Apple iPhone 17 Air Will Be Only 6 mm
-
Life2 weeks agoSony Releases Its New Flagship Camera Alpha 1 II
-
Cinema and Art2 weeks agoNew Trailer of a Minecraft Movie Released!
-
Software2 weeks agoVoice Message Conversion from WhatsApp: Now You Can Read Your Messages!
-
Software2 weeks agoiOS 18.1.1 Update for iPhone Released What’s New?
