Now, browsers are about to get a little more interesting.
Generative AI is coming to every web browser. Google Chrome, Microsoft Edge, and several other browsers are using AI to summarize web documents and improve search efficiency, changing the way people interact with the web, said Ritu Jyoti, group vice president of AI infrastructure at IDC.
“Their goal is to reinvent web search so you don’t have to open dozens of browser tabs and sift through them one by one to find the information you need,” Joti said.
Since the dawn of the web, the way to find content has always been to browse a list of links. Now, generative AI-based chatbots/virtual agents can understand questions posed in plain language and provide direct answers. AI assistants can also help with writing and polishing text, summarizing web pages, and translating them into other languages.
A new chip that builds AI PCs
All of this is possible because of advances made by chipmakers. Both Intel and AMD have announced new generations of desktop, laptop, and mobile processors with neural network coprocessors built in. Apple has its own custom chips, and Qualcomm is eyeing the desktop/client space with its Snapdragon X processor lineup. The idea is that PCs with neural processing units (NPUs) could handle some generative AI processing tasks, like interacting with chatbots like ChatGPT, right on the PC itself. Most AI workloads now happen in the cloud.
“The NPU performance is around 40 TOPS (trillion operations per second), which is more than enough for ChatGPT inference purposes. Previously, it was underperforming at around 20 TOPS,” said Rob Enderle, senior analyst at consulting firm Enderle Group. The current top-of-the-line Intel processor, Meteor Lake generation, delivers about 34 TOPS across CPU, GPU, and NPU, and the upcoming Luna Lake generation will deliver up to 50 TOPS. AMD Ryzen Pro 8000 delivers 39 TOPS, and Qualcomm Snapdragon X Elite delivers 45 TOPS.
Enderle noted that such processors are just starting to hit the market, and that most AI workloads are still cloud-based, so “for this hardware to reach the market it’s supposed to, we’ll first need to see PCs with these chips deployed in large numbers.”
However, browser developers are moving quickly to address AI capabilities, and some will actually take advantage of the new hardware. Most of the AI capabilities currently offered by browsers can be implemented without a neural network coprocessor. However, for more ambitious AI projects to be implemented, there will first need to be a sufficient installed base of new processors capable of performing functions such as LLM processing.
AI features applied to browsers
Let’s take a look at their AI-related activities on browsers big and small, desktop and mobile. Some companies have not yet added AI features to mobile browsers, citing the lack of performance on the device, but others have enabled AI features to work on the device.
Google Chrome
AI features that Google is working on for Chrome include Tab Organizer, which creates tab groups based on the currently open tabs. This feature helps users manage multiple tasks at once by grouping related tabs together. Chrome is also adding the ability to create custom themes based on themes, moods, visual styles, and colors chosen by the user. Finally, there is also a “Help me write” feature that helps draft web content, such as online reviews or inquiries.
These features are being rolled out gradually, with the tab organizer and AI-generated themes introduced in Chrome version M121, and the writing assistant in M122. They are currently available in the US via Chrome on Mac and Windows PCs. Users will need to enable them manually on the ‘Experimental AI’ page in Chrome’s settings.
Microsoft Edge
Copilot is Microsoft’s latest AI assistant project, a replacement for Cortana, which is far behind in terms of functionality. Copilot is integrated into Microsoft’s vast suite of products, including Windows, Bing, Microsoft 365 apps, and the Chrome-based Edge browser.
Microsoft Edge’s CopilotIt provides three core functions:
- chatting : It supports natural language interaction, allowing users to interact with AI through conversations and receive web page summaries or answers to questions.
- writing : AI-based writing assistants help with a range of writing tasks, such as drafting emails or writing blog content.
- Insight : This tool provides contextually relevant information based on the web page the user is currently browsing, presenting additional perspectives and related data.
To use Copilot, you must first install the latest version of Microsoft Edge. Clicking the Copilot icon in the upper right corner of the browser window will open the Copilot sidebar. You can type your question in the window below the sidebar, use one of the default prompts (for example, ‘Create page summary’), or click the Writing tab to get help writing.
Brave
Brave is a browser based on Chrome core technology that focuses on privacy and blocking user tracking. Brave’s Leo AIIt helps with tasks such as generating summaries of web pages or videos, answering questions about content, creating new content, translating pages, analyzing information, and editing text. It also supports various production processes such as writing letters, emails, articles, and blogs.
Brave says chats with Leo are private, anonymous, and secure. Leo doesn’t record conversations or use them to train models, and it can be used without a login or account. Leo integrates into Brave Search, adding AI-enhanced responses to search results. Brave claims this provides more accurate and up-to-date information. Leo is based on several AI models, including Mixtral, Claude, and Llama. Unlike Chrome and Edge’s AI features, Leo is also available on Android and iOS. It appears that the mobile app will have all the same features as the desktop version.
To access Leo on your desktop, open the Brave browser, click on the three lines in the top right corner, then select More tools → Show sidebar → Always. A sidebar will open on the right side of your browser. Clicking on the Leo button in the sidebar will open a chat window.
On the mobile app, open the Brave browser app, tap the three-dot icon, and then select Leo from the menu options.
opera
Opera’s integrated AI assistant, first implemented in mid-2023 AriaAt the time, it was very bold. At the time, most generative AIs, including OpenAI’s GPT, were trained using data that was more than two years old, but Aria combined GPT-based technology with the latest data from the web to provide more up-to-date information in its responses. However, recent generative AI chatbots also use the latest data.
Like other AI programs, Aria helps with tasks such as explaining complex information, translating content, generating code, and answering product-related questions. Aria supports over 50 languages, is available in over 180 countries, and runs on Windows, macOS, Linux, and Android platforms.
Aria’s Reuse and Rephrase feature allows users to easily refine and add to previous responses without having to retype the entire entry. The My Style feature allows you to teach Aria to write in your own style by providing writing samples in formal, neutral, and informal tones.
To access Aria, open the Opera browser and type in the address bar: opera:settingsEnter . Scroll down to the ‘Advanced’ section and find ‘AI Services’. Toggle this section to enable Aria in the sidebar.
Mozilla Firefox
Mozilla is taking a somewhat different approach to AI in Firefox: Users choose their preferred search engine in the same way. Choose your preferred AI chatbotYou can. The options Mozilla suggests include ChatGPT, Google Gemini, HuggingChat, and Le Chat Mistral, with more to come. The feature is currently in an experimental version of Firefox, intended for testing new features. Nightlyis provided in .
In addition, Mozilla is working on adding AI features that “solve real-world problems, respect users’ privacy, and give them real choices.” Mozilla is exploring ways to further enhance the browsing experience using local, on-device AI models. One feature coming next quarter is AI-generated alternative text for images embedded in PDFs, which will improve accessibility for people with visual and learning disabilities.
Mozilla also said it will develop better alternatives to AI, which has become untrustworthy as many chatbots are polluted with misinformation and used for fraud, including deepfakes. To that end, Mozilla is focusing on finding trustworthy open-source AI opportunities and building businesses around them. Mozilla.aiInvested $30 million (about 41.52 billion won) in a startup called .
Apple Safari
Safari 18 browser will be included in macOS Sequoia and iOS 18, which will be released later this year. Safari 18 is said to be incorporating AI into the browser’s features, and some changes are currently underway. It’s still early days, but developers testing the browser say it includes a new “Intelligent Search” feature. This feature appears to use Apple’s Ajax language learning model to identify topics and key phrases and generate page summaries.
But the bigger news from Apple is Apple Intelligence, which was unveiled at WWDC recently. Apple Intelligence is a suite of AI tools designed to work across Apple’s entire product line, for tasks like rewriting, proofreading, and summarizing text, and creating and editing images. The conference also announced that Siri, Apple’s virtual assistant, will be enhanced with ChatGPT integration. It’s still early days and months away from launch, so it’s hard to say exactly what Apple Intelligence will do for Safari and the rest of the operating system. We’ll know more in the fall, when new versions of macOS, iOS, and Safari are released.
One thing is certain: no matter which browser you use, you will soon be able to get help from an AI assistant.
editor@itworld.co.kr
Source: www.itworld.co.kr
