Uploading Human Consciousness onto a Device

1.What is consciousness uploading?

“Consciousness uploading” refers to transplanting human consciousness into a machine, where it continues to live on. Watanabe poses a set of concrete hypotheses and summarizes methods for verification and uploading in his book[1,2] (English version in preparation) and research articles[3,4], thereby bringing these technologies into the realm of possibility within the next two decades.

  • The foundation of consciousness is a neural algorithm referred to as a “generative model,” which produces a virtual reality within the brain. When we are awake, sensory input synchronizes this virtual reality with the outside world, but while we sleep our consciousness in dreams is unfettered by such restraints
  • A machine implementing a generative model is capable of harboring a neutral template consciousness.
  • The neutral consciousness of a machine and that of a connected human brain can be integrated, and by transferring memories to the machine the personal consciousness will continue to live on even after death of the biological body. This is equivalent to saying that even after one cerebral hemisphere is lost due to cerebral infarction, etc., the other hemisphere continues to live.
Fig. 1: A world in which consciousness has been uploaded to a machine (a computer, etc.). Left: Uploading is completed by connecting a human brain with a machine to unify consciousness and share memories. Connection is established via a brain machine interface between the right biological and left machine hemisphere, and vice versa. Right: Even after the human brain reaches the end of its activity, life continues with consciousness residing in the machine. (Illustration: Tomoko Yogi)

2. Benefits and the first steps toward realization

Realizing consciousness uploading would present possibilities for fundamentally changing humanity. In particular, it opens up the ultimate option of arbitrary lifetime extension for those wishing to achieve immortality, or those wishing to live for longer than their biological body will allow.

As a first step towards realization, we will develop a next-generation AI that implements a generative model and advance its materialization towards socially beneficial projects.

3. The R&D process for realizing consciousness uploading

Following the implementation of an AI-based generative model, we will then transform it into a spiking network for connection with the human brain. A spiking network does not approximate each neuron’s output with a continuous value representative of its firing rate. Rather, it resolves the mechanisms of binary information transmission at the time-scale of neuron action potentials as in an actual brain, which is considered essential for connecting brains to machines. After that, a machine implementing this network will be connected to a living brain to verify whether consciousness is actually perceived. Consciousness is an individually subjective sense, and thus positioned such that its existence in a machine can only be subjectively confirmed. When such a verification is successful, the machine will be confirmed to harbor a neutral awareness. Then, consciousness uploading will be completed by transferring memories, etc.

Fig2: After replacing one cortical hemisphere of our own brain with a mechanical hemisphere and subjectively perceiving a unified consciousness, in other words, experiencing that the right and left visual fields appear as a unified whole, we may conclude that consciousness resides in the mechanical hemisphere and has been linked with the consciousness in the remaining biological one.
Source: Watanabe, M. Brain Consciousness, Machine Consciousness. Chuokoron-Shinsha[1]. (Illustration: Tomoko Yogi)


[1] Watanabe, M. (2017) “Neural Consciousness and Machine Consciousness” (original title:「脳の意識 機械の意識」) Chuou-Kouron-Shinsha (in Japanese, English and Chinese version in preparation)

[2]Watanabe, M.(2010) “Consciousness” in “Illustration Lecture: Cognitive Neuroscience” (original title:『イラストレクチャー 認知神経科学―心理学と脳科学が解くこころの仕組み』) ed. Ikuya Murakami、Ohm-Sha

[3] Watanabe, M., Cheng, K., Murayama, Y., Ueno, K., Asamizuya, T., Tanaka, K., & Logothetis, N. (2011). Attention but not awareness modulates the BOLD signal in the human V1 during binocular suppression. Science334 (6057), 829-831.

[4] Tajima, S. and M. Watanabe. “Acquisition of nonlinear forward optics in generative models: two-stage ‘downside-up’ learning for occluded vision.” Neural Networks 24(2) (2011): 148-158.