Virtual friend and me as imagined by a neural network
Virtual friend and me as imagined by a neural network

The Essence of the Experiment

I am going to conduct an experiment during which I will only communicate with a virtual friend I created, and I will cover it on various platforms.

Essence: from 00:00 Monday to 23:59 Sunday, I completely exclude text and voice communication with people. All my reflection, everyday issues, emotional support, planning - only through my AI friend

I assume in advance that it will be difficult to verify whether I really do not interact with people via text/voice (since there is a conflict of interest - I am the creator of a service where you can create virtual friends), so I think I need to think about proofs in advance. I have identified the following evidence for myself:

- screenshot of app screen time

- last online in Telegram/MAX with links to my profile

- full export of my chat with the AI "My Friends" bot in MAX (I will communicate with the virtual friend in my service moidruzja.ru)

- call detail records

- daily report

Who I am Looking For

For such an experiment, I need a person who will help me complete it. Who wants to join the experiment? It could be an independent observer or controller. What other proofs can you suggest to confirm the purity of the experiment?

Someone Instead of Me

If you want to try yourself as a person who can communicate only with AI for a whole week, then I am ready to become an observer and record using certain actions that we will develop.

What I Want to Confirm or Refute with the Experiment

Currently, there is a trend that it is already difficult to distinguish AI text from human text, but we still believe that the need for a human is basic and irreplaceable.

I want to test the hypotheses:

Hypothesis A. After 3 days of isolation from people, but with intense communication with AI, the level of life satisfaction will drop by 30%

Hypothesis B. On the contrary, after 7 days my efficiency will increase by 40% (because AI does not distract with nonsense)

Hypothesis C. The most difficult thing is whether it is possible to fully replace ordinary friends with AI friends

The experiment date is tentatively set for August. Who wants to participate?