- Published on
Humanely AI, a hope
- Lorenzo Pieri
Humanely AI, a hope
I started this blog post like many others. Opened a few pages, sprinkled the idea, wrote a basic input on a particular Notion page and collected more resources to gather more insights and knowledge about what I wanted to write.
The thing is, today I could have just wrote the prompting message and deliver it to what’s been on everyone’s homepage lately: chat-gpt3, also known as the AI that will steal our jobs, kill us or use us as milking cows until we are too old of age.
Probably went a little bit overhead with this last naming convention but one should know that even the strongest of mind and belief are faltering a little bit once caught in the endearing endeavours of this mighty technology.
I too have succumbed to the flattering promises of a lightweight solution and ease of prompt. I too have had the fun to input the most hilarious, or distasteful, or misogynistic, or philanthropic, or racist, or exquisite questions or remarks to this astonishing piece of ars tecnica. I have tried many ways to trick the AI to give answers that were far from perfect. Some of them far from decent. It's an AI.
It is far from perfect and sometimes it can help us peek into the known and unknown worlds of human knowledge and interactions. The language model of chatgpt-3 is built upon human literature of all kinds, after all. Be it blogs, social media platforms, books, reviews, papers, conversations, movie dialogues, audios. Whatever it is, it moulded this AI into what it is. Depending on the inputs there are ways to collect views that are all but nice and human.
I am sure you might have seen code generation where the examples are based upon parameters that might discriminate through the output. I will leave to the user the search for such examples as I find them very sad and disheartening given that these answers are based upon training that was “cooked” by humans.
Despite the many uninteresting articles that were released into the wild during this last month I still am very fond and entertained by this new powerful ally we have now at everyone’s hands. This led to questions, that led to more questions which featured questions that ended in questions. I guess there is no good answer to this.
The starting question is:
Is humanity good enough for AI?
Usually it goes the other way around but this time I am wondering whether or not we deserve as specie to be able to use such a powerful and versatile instrument.
As we learnt to use primeval tools, we started making weapons. We could have ended at hunting weapons, we didn’t.
As we learnt to use fire, we searched for ways to capture its essence. We could have warmed our houses and yet we went in many other places with this ability tho, not all of them great.
As we learnt to use the internet, we tinkered with its many secrets. We could have built a new era for humans but what we created, amongst many things, were phishing, smishing, vishing and hacking in general.
Don’t get me wrong. Hacking is great if you think of yourself as an actual hacker. Trying to f*%k with people’s life and making them miserable is not what a hacker does. That’s another thing.
One might argue that with any invention or discovery there’s going to be blacks and whites. There’s going to be middle grounds, the famous grey areas.
I concur and yet this is exactly what makes me wonder whether or not humans are deserving of such powerful instruments. History luckily has nice ways to remind us of how stupid and colossally outrageous we can be as a specie. Think Atomic Bombs. Think Nuclear Bombs. Think many other things. Do we have to put “Bombs” anywhere we invent? I hope not.
But I am digressing from our question. Is humanity good enough for AI? I think that we can be, with the right attitude.
What is possible?
I am no scientist of the Artificial Intelligence field. I’d love to learn more about data science and algorithms of the sort in general but for now I will write about how I see this moving forward as a human being and not as a professional.
Let’s make a small list of the most sophisticated public AIs that are available right now for everyone to use:
- chatgpt-3, we all know
- midjourney, an AI that can generate images based on user prompt
- deepfake, not an AI but rather a technique to generate fake imagery that looks real
- DALL-E, another image generator
There are many more AIs that are becoming public each day and some models are also open sourced and available at anytime to anyone. This is fantastic but also scary.
Look for VALL-E, a Speech Synthesizer AI that can generate real human voice. That's not scary at all.
Depending on the usage one could do a great amount of damages to a great amount of people and companies.
Let’s focus on the brighter side
I started this article with a nuance of bitterness and without trust in the human spirit but in reality I believe that we are fundamentally good, at least in the higher percentage of population on the planet. We cannot control what’s good and what’s bad as it is perceived directly by each and everyone of us and it’s strongly dependent on so many factors that would make it not ideal for a calculation.
Let’s pretend that everyone is good for a moment tho, at least for what generally is thought of as “a good human being”.
What would you do with this fast evolution of the technology?
Humans are fundamentally made to connect
I’d like to think that this evolution of AI would enable us humans to become even more connected, allowing us to exchange thoughts, processes and feelings with beings and other humans that are not necessarily present.
I imagine an AI that will be able to take inputs from literature, sounds, voice files, recordings, videos, images and so much more. These inputs would then be used to generate an entity that could resemble a person not only in imagery but also in character, personality and knowledge.
I imagine the possibility to talk with the greatest minds of past times or with someone who is no longer with us but somehow their memory still exists.
I am not talking about immortality of the person or immortality of the mind. I am far too scared of such a concept and I find it unnatural, unrealistic and also very wrong.
I am talking about shared knowledge and history, emotions that are never to be lost in time, voices that can be heard at any moment.
I imagine that a good AI will be as good as the collective knowledge and feelings of humanity.
What about you? What do you think?