{"id":17037,"date":"2022-12-06T13:36:27","date_gmt":"2022-12-06T13:36:27","guid":{"rendered":"https:\/\/uxmag.com\/?p=17037"},"modified":"2022-12-07T15:47:53","modified_gmt":"2022-12-07T15:47:53","slug":"conscious-ai-models","status":"publish","type":"post","link":"https:\/\/uxmag.com\/articles\/conscious-ai-models","title":{"rendered":"Conscious AI models?"},"content":{"rendered":"\n

A few days ago, the Internet was taken by countless tweets, posts, and articles about Google’s LaMDA AI being conscious (or sentient) based on a conversation it had with an engineer. If you want, you can read it here<\/a>.<\/p>\n\n\n\n

If you read it, you will also realize that it surely looks like a dialog between two people. But, appearances can be deceiving\u2026<\/p>\n\n\n\n

What IS LaMDA?<\/h3>\n\n\n\n

The name stands for “Language Model for Dialogue Applications”. It’s yet another massive language model trained by Big Tech to chat with users, but it’s not even the latest development. Google’s blog has an entry from more than one year ago called “LaMDA: our breakthrough conversation technology”<\/a>. It’s a model built using Transformers<\/em>, a popular architecture used in language models. Transformers are simple, yet powerful, and their power comes from their sheer size.<\/p>\n\n\n\n

LaMDA has 137 BILLION parameters, and it was trained on 1.5 TRILLION words (for more implementation details, check this post<\/a>).<\/em><\/p><\/blockquote>\n\n\n\n

To put it in perspective, and give away my age, the Encyclopaedia Britannica has only 40 million words in it. So, LaMDA had access to the equivalent of 37,500 times more content than the most iconic encyclopaedia.<\/p>\n\n\n\n

In other words, this model had access to pretty much any kind of dialog that has ever been recorded<\/em> (in English, that is), and it had access to pretty much every piece of information and knowledge produced by mankind<\/em>. Moreover, once the model is trained, it will forever “remember<\/em>” everything it “read<\/em>“.<\/p>\n\n\n\n

It’s impressive? Sure! It’s an amazing feat of engineering? Of course! Does it produce human-like dialog? Yes, it does.<\/p>\n\n\n\n

But, is it conscious, or sentient? Not really\u2026<\/p>\n\n\n\n

“Why not?! If it looks like a duck, swims like a duck, and quacks like a duck, then it probably is a duck, right?”<\/em><\/p><\/blockquote>\n\n\n\n

Yes, it is probably a duck. But that’s because we know what a duck is (and LaMDA knows it too, or at least it can talk about ducks just like you and me).<\/p>\n\n\n\n

What IS consciousness?<\/h4>\n\n\n\n

Unfortunately, the definition of consciousness is not as obvious as the definition of a duck. And, before the late 80s, no one was even trying to define it. The “C-word<\/em>” \u2014 consciousness \u2014 was banned from the scientific discourse.<\/p>\n\n\n\n

Luckily, the scientific community forged ahead, and nowadays it has a much better understanding of consciousness. But the everyday usage of the term consciousness hasn’t changed, and it still covers a lot of different<\/em> and complex<\/em>phenomena.<\/p>\n\n\n\n

What does \u201cbeing conscious\u201d usually mean when one uses the term? For one, being awake \u2014 \u201cshe remained conscious after the crash<\/em>\u201d. Alternatively, being aware \u2014 \u201che is conscious of his faults<\/em>\u201d.<\/p>\n\n\n\n

In order to properly study a topic, though, it is required to properly define<\/strong> the object of study. And that’s what Stanislas Dehaene does in his book, “Consciousness and the Brain<\/strong>“, from which I am drawing the majority of the ideas in this section. The author distinguishes three concepts:<\/p>\n\n\n\n

  1. vigilance<\/strong>: the state of wakefulness;<\/li>
  2. attention<\/strong>: “focusing mental resources on a specific piece of information<\/em>“;<\/li>
  3. conscious access<\/strong>: “the fact that some of the attended information eventually enters our awareness and becomes <\/em>reportable<\/em><\/strong> to others<\/em>” (highlight is mine).<\/li><\/ol>\n\n\n\n

    Hold this thought: with conscious access, the information becomes reportable. We’ll get back to this soon!<\/em><\/p><\/blockquote>\n\n\n\n

    For Dehaene, vigilance and attention are required, but not sufficient, and only conscious access qualify as genuine consciousness.<\/p>\n\n\n\n

    Conscious Access<\/h4>\n\n\n\n

    It looks trivial, right? You see something, say, a flower, and you instantly become aware of its properties, color, smell, shape, etc.<\/p>\n\n\n\n

    So, you can safely say that you’re aware of everything your eyes see, right?<\/p>\n\n\n\n

    Please watch the short video below (and don\u2019t scroll down past the video otherwise you\u2019ll spoil the answer!):<\/p>\n\n\n