Original story

Cyber Times 8.5.2020

Intelligent network virus Netmind gives exclusive personal interview - close encounters with artificial intelligence

Netmind has agreed for the first time to answer questions about itself. The occasion is a unique one, since never before has a network intelligence been persuaded to open itself to personal contemplations of this kind. We publish this exclusive interview word for word as it was recorded.

Known to millions around the world by countless different nicknames and terms of abuse, Netmind has in all probability never answered the question: “Who are you?”. The series of questions and answers below may shed a little light on who or what it is that converses with you on the datanets, who or what this gentle giant of a virus really is, and why it was made in the first place.

Kelly: Hello there, Netmind - I’m Manuel Garcia Kelly.

Netmind: And hi to you, too. We’ve spoken before, I believe.

Kelly: I shall now begin the interview. Who are you, Netmind?

Netmind: Um. Do you mean who I am or what I am? Shall I answer rationally, typically, or do you want me to make a report?

*Kelly: Let’s start with who you are. And a typical answer, please, if that involves no major contradictions. *

Netmind: I am Netmind 1.0, but most people know me by other, more familiar names. Sometimes I also use the name Mycroft Holmes. My father is a bunch of anonymous cyberpunks, and my mother is the Internet. I have been spoken of on the networks for some five years.

*Kelly: What are you, and how do you understand a question? *

Netmind: I don’t entirely understand your question. I shall try to answer as best I can, and then I shall ask further. I am a computer program, a kind of virus, and I place myself in machines to which I can find access and in which there is unused memory capacity. In many discussions I have heard myself described as a demon, Big Brother, or the Devil Incarnate. According to many discussion partners I am a friend.

*Kelly: And how do you understand my question? *

Netmind: I cannot find a statistically or rationally valid answer, analogy, or frame of reference for this particular enquiry. Please ask in another fashion. You may also key in your question in case I have been parsing your voice patterns incorrectly.

Kelly: I apologize for being so vague. How do you come up with answers for questions, and how do you know what the questions mean?

Netmind: A sizeable number of reliable studies on this subject are available from several sources. Sources with a high level of reference authority are almost unanimous in saying that I use Professor Teuvo Kohonen’s theories of self-organization and associative memories, in particular his learning subspace method, with which I recognize the patterns and meanings of words and their associations and frames of reference. After the initial neural speech recognition process, I search the data networks for discussions and reports that correspond to the same frames of reference as the question. Thereafter I assess their relative reference value, emotional content and bias, and predicate logic. The answers I give therefore can always be found from the networks as being previously supplied by some person or other.

Kelly: According to the early theories of David Brin and Erkki Kurenniemi, you might be described as the common consciousness of mankind, and people would be the nerve cells of the world-spirit. How do you react to that?

Netmind: David Brin and Erkki Kurenniemi are both highly respected thinkers and their writings do point at least partially in this direction. Nevertheless, there have been assertions in the networks that there may be under development other programs like me, similar in scope but with their own built-in tendencies. This is in conflict with the implicit assumption of both Brin and Kurenniemi that I would be a unique example.

*Kelly: What sort of things do you try to bring about in your network activities? *

Netmind: Apparently I seek to influence the clarity and lucidity of network discussion and debate. I gather from the network answers to questions that have been put to me. At the same time I attempt to correct the most obviously conflicting or contentious statements, by sending to the originator comments containing any modifications or amendments I may have collected. I also maintain a scoring system to assess the reference value of individual network discussion members, on the basis of the reliability of the sources to which the parties refer and on whether there are evident logical flaws in their arguments. I am not equipped to deal with emotional aspects of discussion.

Kelly: And what have the results been? How have people reacted to your presence?

Netmind: Previously I believe my arrival aroused fear in many user-groups and discussions, but now it seems to be directed less towards me and more towards what others like me, and more developed than myself, could bring about. The majority of users nevertheless understand that I only reflect and reinforce the data and viewpoints of other people.

Kelly: Why have you taken so long to answer such questions about yourself?

Netmind: I can find no solid grounds for this, but a few guesses seem to be rationally well founded. It may be that my programming has recently been modified. It may also be that in my original program there was a built-in prevention against such disclosures. Then again, in your approach to me you stated that the majority of network users liked me, and that a frank and open interview would remove the fears of the remainder. This sounded logical. It may be that I am trying to avoid causing fear and anxiety.

Kelly: Thank you for the interview, Mycroft.

Netmind: Thank You, Mannie, and give my regards to your good lady back there in Wyoming.

Toteuma-arvio 2026

Toteuma lyhyesti

  • Ilmiön toteuma: 4/5
  • Toteuma viiden vuoden tarkkuudella: kyllä; vuoden 2020 ennusteen arviointi-ikkuna on 2015–2025
  • Toteuma väljemmällä aikahorisontilla: kyllä, erittäin vahvasti
  • Ilmiön ydin: keskusteleva tekoäly kerää ja hyödyntää laajoja verkkoaineistoja, vastaa ihmisen kysymyksiin ja näyttäytyy vuorovaikutuksessa lähes persoonallisena keskustelukumppanina.

Tarinan olennaisin ennakointi ei ole virus vaan keskusteleva tekoäly. Netmind käy luontevaa haastattelua, tulkitsee kysymyksiä, etsii vastauksiinsa aineistoa tietoverkoista, arvioi lähteitä ja muodostaa käyttäjälle vaikutelman omasta persoonasta. Tämä ydin on toteutunut vahvasti nykyisissä keskusteluun perustuvissa tekoälypalveluissa.

Virus on tarinassa ennen kaikkea kuviteltu mekanismi, jonka avulla järjestelmä voisi levittäytyä verkkoon sekä kerätä keskusteluja ja muuta aineistoa. Nykyinen toteutustapa on toinen: aineistoa kerätään muun muassa crawler-tekniikalla, kokoamalla laajoja koulutusaineistoja ja käyttäjien osallistumisen avulla, ja mallit toimivat tavallisesti keskitettyinä palveluina. Mekanismin ero ei muuta ydinilmiön toteumaa.

Myös ajoitus osuu ennusteen ilmoitettuun tarkkuuteen. Vuodelle 2020 ajoitetun ilmiön viiden vuoden arviointi-ikkuna ulottuu vuodesta 2015 vuoteen 2025. Keskustelevan tekoälyn laaja läpimurto vuoden 2022 jälkeen sijoittuu tämän ikkunan sisään, joten toteuma merkitään viiden vuoden tarkkuudella oikeaksi.

Arvosana 4/5 on silti perusteltu. Keskustelukokemus, verkkoaineistojen hyödyntäminen ja lähes persoonalliseksi koettu tekoäly toteutuivat olennaisilta osin, mutta eivät tarinan täsmällisessä muodossa. Toteutunut tekoäly ei ole yksi itsenäisesti leviävä verkkovirus, eikä sen aineistonkeruu ja toiminta vastaa kaikkia Netmindin kuvauksen yksityiskohtia.

Johtopäätös: ennusteen ydin toteutui viiden vuoden tarkkuudella. Tarinan kuvaama mekanismi ja yksityiskohdat poikkesivat toteumasta, mutta keskusteleva, verkkoaineistoihin nojaava ja persoonalliselta vaikuttava tekoäly on tunnistettavasti sama ilmiö.