Are You Talking To Me?

Some thoughts about communicating via the Internet.

* * *

A newly acquired old camera had a small problem, so I went for some help to the Photrio site which caters to film photographers.

My request:

I recently acquired a Yashica YE. Everything works well except the frame counter which does not advance.
I would appreciate any insights on getting into the counter mechanism and fixing the problem.

Here was the first response I got:

Could be a number of things. You will have to open the camera and examine the counting mechanism to see what is wrong.

Here is a case where a little plastic piece broke, causing film counter failure...

The lack of specificity in the reply along with veering off into a lengthy explanation of an unrelated problem signaled that the response was AI-generated.  Someone had clearly just submitted my question to a chatbot and posted the result without further elaboration.

Another indication I have seen recently of AI intrusion is at Slashdot, a site where people post summaries of reports about interesting technical developments.  Many of the summaries have gotten  longer, which I think is the result of AI use.  Composing a good summary requires some time and effort which is saved by the AI intervention.  So, in this case, probably a useful implementation.

On the other hand, one has to wonder where all this is heading.  It seems possible that Slashdot and similar websites could ultimately be managed with virtually no human intervention.

I asked the DuckDuckGo AI what percentage of postings at Facebook were likely produced by AI and was informed that about 40% were not from humans.  The amount of images produced by AI on the Internet looks to be around 70%.

Since it is apparent that AI postings on social media and in the academic world are fairly easy to detect I am wondering if website operators might be asked to provide AI-free versions of what is offered, and if that is something that site visitors might find attractive.  It certainly has some appeal for me.

In fact, I have found that some success in avoiding AI is within my grasp without any outside intervention.  Much of the AI content one comes across is generated by algorithms that tailor what is presented based on AI-perceived preferences and personal characteristics which AI monitors have detected and accumulated about each of us.  So avoiding detection of your identify when interacting with a website is a step toward sought-after anonymity.

One way to achieve at least partial anonymity is to right-click on links and select the option to "Open in New Private Window". Give that a try, for instance, with accessing Google News and you will get a much pared down version of the daily news without all the click-bait fluff based on your personal record which has been assembled by the AI algorithms.

The "Private Window", of course, is still a window and some revealing information will get through including your geographic location and browser choice.  So, if you really want more anonymity you will need to take further steps like adjusting your browser for maximum security, or even using a Virtual Private Network (VPN) for accessing the Net.

The problem with those privacy measures is that they can quickly begin to interfere with the functionality and usefulness of the user's experience.  Such measures, if used widely, will also affect the bottom line of the big site owners for whom your identity is a commodity.