‘Agentic browsers’

[Mæstro]

Through a joke which a friend had shared in a chatroom, I learnt about so-called agentic browsers:

But, more recently, we are seeing the rise of agentic browsers, which are a new type of web browser powered by artificial intelligence, designed to do much more than just display websites. These browsers are designed to actually take over entire workflows, executing complex multi-step tasks with little or no user intervention, meaning they can actually use and interact with sites to carry out tasks for the user, almost like having an online assistant. Instead of waiting for clicks and manual instructions, agentic browsers can navigate web pages, fill out forms, make purchases, or book appointments on their own, based on what the user wants to accomplish.

For example, when you tell your agentic browser, “Find the cheapest flight to Paris next month and book it,” the browser will do all the research, compare prices, fill out passenger details, and complete the booking without any extra steps or manual effort—provided it has all the necessary details of course, which are part of the prompts the user feeds the agentic browser.

The possible outcomes of permitting a browser to operate autonomously like this are easy to imagine, but i had not imagined that the same article would suggest that among these are the equivalent of a Neopets cookie grabber, but for one’s real bank account. (I hope others here played enough back in the day to catch this allusion.) It is obvious Pale Moon will never implement this sort of vulnerability, but is there any actual word about the monopolist powers encouraging this? The most I could find is, hilariously enough, Norton hawking this technology.

[Night Wing]

I would never use an agentic browser because I would never trust it. In other words, my trust is never given. My trust has to be earned.

[Moonchild]

I don't think this will have a good time working unless the web as a whole becomes "agentic-friendly" in some way.
e.g. How would it deal with ads, scams, many different website constructions that are already throwing standard browsers for a loop, etc.?

Not to mention the fact that current "AI agents" can already be fooled by taking content responses as command input (big no-no...); now imagine that while having given it agency (pun intended) over filling in forms, making payments, etc.

[RealityRipple]

I don't think this will have a good time working unless the web as a whole becomes "agentic-friendly" in some way.
e.g. How would it deal with ads, scams, many different website constructions that are already throwing standard browsers for a loop, etc.?

Not to mention the fact that current "AI agents" can already be fooled by taking content responses as command input (big no-no...); now imagine that while having given it agency (pun intended) over filling in forms, making payments, etc.

Albania's PM literally just appointed their "government document ai helper" to their cabinet, so I'd say that bridge is being set alight as we speak.

[Night Wing]

Albania's PM literally just appointed their "government document ai helper" to their cabinet, so I'd say that bridge is being set alight as we speak.

I'm retired from the maritime industry. With your remark with the word "bridge" in relation to ships at sea; with these types of browsers, in the maritime industry most of us would say "it looks like the bridge on the Titanic".

[UCyborg]

Am I crazy for thinking the future predicted by The Terminator would be more interesting than the AI bullshit that emerged now?

[Night Wing]

@ UCyborg

Just imagine if the machines of AI today become "self aware" like the machines in the Terminator did.

Just "food for thought".

[UCyborg]

My mind is a dark place and my sanity might be on decline.