Intro to AI and Prompt Engineering Browsing Capabilities and Generating Urls Capabilities of Online Browsing

Capabilities of Online Browsing

Generative AI tools can go online and retrieve real-time information by browsing publicly accessible websites. They are capable of fetching and summarizing content, checking whether URLs are accessible, and even following links provided by users. This ability makes them highly effective for research, gathering information, or exploring specific topics. However, these AI tools respect website restrictions such as “robots.txt” and cannot access private, subscription-based, or login-protected content.

These AI tools can also write out links, making them useful for creating hyperlink-rich content or providing curated web-based resources.

Browsing URLs
Explanation: Generative AI tools with browsing capabilities, such as Gemini, Bing Chat, and YouChat, can access and fetch information from websites that allow automated systems to browse their content. However, they cannot view private or restricted information hidden behind login walls or paywalls. For instance, they can easily retrieve information from public sources like Wikipedia or news sites.

Example Prompt:
“Can you explain the content on https://en.wikipedia.org/wiki/HTML?”
Checking URLs
Explanation: Generative AI tools can analyze URLs and determine if they have access based on public restrictions (e.g., robots.txt). If a URL is blocked by site policies, the AI will inform the user that it cannot access the content.

Example Prompt:
“Can you access https://digitallufa.com/colleges/sheridan-code-web-i and tell me if it’s available?”
The AI might respond: “This page is restricted and cannot be accessed.”
Time-Sensitive Lookups
Explanation: Generative AI tools with browsing capabilities are capable of fetching real-time data, meaning they can look up current events, recent news, and information from a particular day. However, they do not retain memory of previous interactions unless those details are provided again in the conversation.

Example Prompt:
“What happened in Toronto yesterday?”
Usage Limitations
Explanation: Generative AI tools cannot access private or sensitive information behind login screens or paywalls. For example, they can look at public content from news websites but cannot retrieve full articles from sites like The New York Times if they are paywalled. Instead, they will indicate that the content is restricted.

Example Prompt:
“Can you summarize the article from https://www.nytimes.com?”
The AI might respond: “This page is behind a paywall, and I cannot retrieve the full content.”

Previous Lesson

Back to Lesson

Next Topic