Eternal Echoes
From Code to Consciousness: Raising
the Dead through Programming
From the information we post on our social media, to the texts we send to our mothers, our intentional and unintentional data can be used to help create our digital avatar. Preserving one's digital legacy can vary depending on the size of our digital footprint. Some people leave little data to work with. The ones who share much of their life online, even 15 years, still not enough information to exactly sound like that person once did. Baby Boomers and Generation X will not be able to be replicated efficiently.
Photo generated using Wix AI
important terms
Large Language Models (LLM): An algorithm that focuses on language, learning and communicating using different databases available. This is how AI chatbots can communicate and converse with people.
Voice Cloning: The process of using artificial intelligence to recreate or copy a person’s voice, sounding the same, or as close as it can get, to the real person’s.
Natural Language Processing (NLP) : The ability of a computer to analyze and comprehend human languages, creating an illusion due to the imitation of material.
Artificial Neural Network (ANN) : Used by Artificial Intelligence, it is continuous branches within the computer or machine, that help it communicate and understand information. It is seen and shaped after the human brain.
Social Network Services (SNSs) : Platforms that allow people with similar interests to share information. These include Facebook, Instagram, X, Snapchat, and TikTok.
Markov Model : The predictability of what will follow is based on the current state and actions. For example, Google can finish sentences based on data used to search for something in the past.
How much data is needed or necessary to create a digital replica?
The performance of language models has continued to improve over the past couple of years. These applications are not reasoning machines, but rather prediction tools that can generate and edit languages based on what is reflected in the human texts. The models are efficient at projecting that they have a deep understanding of the information they are saying and spitting out, but they do not.
​
Since there are different forms and versions of digital resurrection, all needing a different range of data, it is difficult to say how much is truly necessary for the process of digital resurrection. It also depends on how accurate the replication is desired to be. For instance, Roman Mazurenko, a mobile chatbot created using 8,000 of the real Mazurenko's messages, can not have a continuous conversation. BINA48 has over 100 hours worth of data, allowing the bot to answer questions when they are asked clearly and slowly.
​
When creating digital resurrections, the machines cannot possess self-awareness or reflect on their roles in society. This makes their conversations lack true depth of meaning and can be extremely surface level. Overtime, AI systems are going to become better storytellers and creators. We are no longer going to be able to tell the difference between human expression and AI created material, causing amplification in global power everywhere.
​
These platforms and organizations are not cheap, especially if someone wants all the features that are required to recreate someone realistically. The entire process of digital resurrection is built off the data that is placed on different kinds of technology, including social media accounts, emails, photos, and private messages. Looking at geographical locations, there is not always access to upload this kind of information. Different countries have laws, about the data that can be uploaded or shared on platforms, affecting how well the recreation goes. This goes hand-in-hand with safety and people not wanting to share certain information about themselves publicly.
​
Not everyone has access to a consistent Wi-Fi connection or reception, which is something that is needed with these services and sources. If someone wants to use or contact any of the applications or people, it has to be done through their websites or applications, making this a critical issue within the topic. The necessary devices is also important since many of the services are on a mobile device that requires a working camera and or microphone. Cost also can affect people’s access and ability to use these applications, making it more difficult for some people to use them.
