The Arc Browser by The Browser Company marked one of the first attempts to comprehensively integrate AI with a web browser.
Last year, The Browser Company browser’s new Arc Max features included ChatGPT integration and tools to organize tabs and downloads. This year, Arc Search, a browser for iPhone, uses AI summarization to provide answers to queries rather than just links.
The desktop Arc Browser is also evolving, with new AI-powered features rolling out this year. Instant Links, for example, lets users bypass search engine results pages to directly reach specific web pages and information. Live Folders offers an automated feed of recommendations related to bookmarked topics.
Perhaps most notably, the upcoming “Arc Explore” feature will actively search for content based on your queries, displaying results and the sources used to find them.
The Browser Company describes Arc Explore as “a tool for automating a browsing journey from end to end.” It aims to replace the multi-step browsing process with a streamlined AI-powered workflow.
So, in light of everything Arc is bringing to the table, is the era of the AI browser finally upon us?
Will the traditional search engine results page (SERP) fall from grace as we rely on AI to answer queries in a more narrative format?
AI browsing could indeed change the face of web browsing, but some, like Tyler Farrar, CISO at Exabeam, are concerned about security risks and other potentially internet-wide impacts.
AI browsing presents security risks
Last year, researchers uncovered security vulnerabilities in Bing’s AI chatbot where compromised ads within the chatbot interface led users to malicious sites, resulting in potential malware downloads.
These kinds of threats might proliferate with newer AI browsers like Arc. Scrutinizing the origin and authenticity of links will become exceptionally important, and threat actors could find methods of exploiting the system.
As Tyler Farrar, CISO at Exabeam, explained, “The Browser Company’s launch of its AI-powered desktop browser Arc represents a significant leap towards intuitive web interaction.”
“As a CISO, the primary concern that arises with such advancements pertains to the security implications. The automation of link retrieval and the bypassing of traditional search engines pose a unique set of challenges, particularly in ensuring the authenticity and security of the links generated by AI.”
Another potential security risk of AI browsers stems from how developers intend to handle sensitive data.
One report found that over half of people’s interactions with chatbots like ChatGPT include personally identifiable information (PII) and other sensitive information. Another found that 11% of employees paste confidential business information into ChatGPT.
Farrar describes how Arc must provide assurances on how they plan to handle any PII input into their system: “Another dimension that warrants scrutiny is user privacy. AI-driven personalization and automated content delivery must be balanced with the assurance that user data is not exploited for unsolicited tracking or profiling.”
“The Browser Company needs to be transparent about their data management practices and provide users with clear controls over their personal information. This is crucial in maintaining trust and upholding privacy standards in an increasingly AI-integrated web environment.”
AI web browsing challenges the internet’s structure
AI browsers essentially curate their results, weakening the link between search results and sources. While source links are provided, it could be tempting to take AI-generated summaries of search results at face value.
Moreover, content restrictions will omit some websites from AI searches, reducing the number of quality sources an AI browser can access. Some 15% of leading websites already block OpenAI’s web crawlers, including the New York Times, the BBC, Amazon, and Wikihow.
If more sites restrict AI-related bots from crawling their links, the pool of information available to AI browsers will reduce, impeding their ability to provide the highest-quality results.
“The implications of Arc for the internet ecosystem are far-reaching,” said Farrar, “While the browser is poised to simplify the multi-step processes associated with traditional web usage, it also alters the dynamics of web traffic and engagement for content creators and marketers. This could potentially disrupt the economic models that current internet infrastructure relies upon.”
When ChatGPT went global, many touted it as the ‘Google killer,’ speculating it’ll suck up Google’s ad revenue, which is the company’s financial mainstay.
This didn’t quite materialize. Google’s ad revenue did drop in 2023 but rebounded to 11% growth in the last quarter.
However, true AI browsers like Arc promise to transform our usual web browsing habits, meaning surfing fewer websites, clicking fewer links, and seeing fewer ads. That could have internet-wide impacts on how web content is produced and disseminated.
In time, there’s a very real possibility that AI will change how people interact with the internet. Navigating the risks while emphasizing the benefits will require diligence on the part of both developers and users.
