Multiple AI agent session browser for Pharo Smalltalk - mumez/pharo-agentic-browser

All-in-one AI aided Smalltalk development container - mumez/pharo-agentic-browser-opencode-docker

Agentic browser at risk Agentic web browsers that use AI tools to autonomously do tasks across various websites for a user could be trained and fooled into phishing attacks. Hackers exploit the AI browsers’ tendency to assert their actions and deploy them against the same model to remove security checks.  According to security expert Shaked Chen, “The AI now operates in real time, inside messy and dynamic pages, while continuously requesting information, making decisions, and narrating its actions along the way. Well, 'narrating' is quite an understatement - It blabbers, and way too much!,” the Hacker News reported. Agentic Blabbering is an AI browser that displays what it sees, thinks, and plans to do next, and what it deems safe or a threat.  Tricking the browsers By hacking the traffic between the AI services on the vendor’s servers and putting it as input to a Generative Adversarial Network (GAN), it made Perplexity’s Comet AI browser fall prey to a phishing attack within four minutes.  The research is based on established tactics such as Scamlexity and VibeScamming, which revealed that vibe-coding platforms and AI browsers can be coerced into generating scam pages and performing malicious tasks via prompt injection.  Attack tactic There is a change in the attack surface as a result of the AI agent managing the tasks without frequent human oversight, meaning that a scammer no longer has to trick a user. Instead, it seeks to deceive the AI model itself.  Chen said, “If you can observe what the agent flags as suspicious, hesitates on, and more importantly, what it thinks and blabbers about the page, you can use that as a training signal.” Chen added that the “scam evolves until the AI Browser reliably walks into the trap another AI set for it." End goal? The aim is to make a “scamming machine” that improves and recreates a phishing page until the agentic browser accepts the commands and carries out the hacker’s command, like putting the victim’s passwords on a malicious web page built for refund scams.  Guardio is concerned about the development, saying that, “This reveals the unfortunate near future we are facing: scams will not just be launched and adjusted in the wild, they will be trained offline, against the exact model millions rely on, until they work flawlessly on first contact.”

  Google appears to be testing a new feature that could significantly change how users browse the web on mobile devices. The company is reportedly experimenting with integrating its AI model, Gemini, directly into Chrome for Android, enabling advanced agentic browsing capabilities within the mobile browser. The development was first highlighted by Leo on X, who shared that Google has begun testing Gemini integration alongside agentic features in Chrome’s Android version. These findings are based on newly discovered references within Chromium, the open-source codebase that forms the foundation of the Chrome browser. Additional insight comes from a Chromium post, where a Google engineer explained the recent increase in Chrome’s binary size. According to the engineer, "Binary size is increased because this change brings in a lot of code to support Chrome Glic, which will be enabled in Chrome Android in the near future," suggesting that the infrastructure needed for Gemini support is already being added. For those unfamiliar, “Glic” is the internal codename used by Google for Gemini within Chrome. While the references do not reveal exactly how Gemini will function inside Chrome for Android, they strongly indicate that Google is actively preparing the feature. The integration could mirror the experience offered by Microsoft Copilot in Edge for Android. In such a setup, users might see a floating Gemini button that allows them to summarize webpages, ask follow-up questions, or request contextual insights without leaving the browser. On desktop platforms, Gemini in Chrome already offers similar functionality by using the content of open tabs to provide contextual assistance. This includes summarizing articles, comparing information across multiple pages, and helping users quickly understand complex topics. However, Gemini’s desktop integration is still not widely available. Users who do have access can launch it using Alt + G on Windows or Ctrl + G on macOS. The potential arrival of Gemini in Chrome for Android could make AI-powered browsing more accessible to a wider audience, especially as mobile devices remain the primary way many users access the internet. Agentic capabilities could help automate common tasks such as researching topics, extracting key points from long articles, or navigating complex websites more efficiently. At present, Google has not confirmed when Gemini will officially roll out to Chrome for Android. However, the appearance of multiple references in Chromium suggests that development is progressing steadily. With Google continuing to expand Gemini across its ecosystem, an official announcement regarding its availability on Android is expected in the near future.

Attached: 1 image Ready to use ChatGPT Atlas? Please don't‼️ ➡️ It can literally be hacked by reading websites ➡️ It reads everything you're logged into: your email, your CRM, your bank account ➡️ "Delete" doesn't mean deleted ➡️ "Incognito" mode isn't private ➡️ GDPR/compliance nightmare More: https://tuta.com/blog/dont-install-atlas-ai-browser-heres-why #OpenAI #ChatGPT #AtlasBrowser #AgenticBrowser

A smarter way to browse

Agentic web browsers are on the rise, learn what they are, how they are different, what industry is saying it's challenges and opportunities.