A backstage entrance into the world of neuroscience and academia.

Ask your AI: Stealing or Starving?

A life full of questions…every day we ask questions and we answer to others. This is how we learn, communicate, evolve and live. Have you ever counted the question marks in a day (here, another one)?

The questions have some weights. Sometimes they are trivial or rhetorical but eventually, they can be crucial and change your complete life. The ‘decision making’ field has been in the focus of the neuroscientists during the last decades. A huge amount of scientific articles answers partly how the animal or human brain works during decisions or uncertainty. We want to get deeper and deeper in the networks and neuronal synapses which are activated after a simple ‘question mark’.
But what we can say about the brain function for those questions and decisions that are ultimately important for our or the life of other? Cognitive neuroscience focuses on the aspects of ‘Moral Judgements’ and ‘Moral actions’. Although, the understanding of neural mechanisms underlying moral decisions has been increased by the recent fMRI studies this topic is still difficult to be studied. The reason is that morality is formed by the society and there is a significant difference between people based on their social background, their reaction to justice, value system or empathy levels. Based on previous work we know that plethora of cortical and subcortical areas, for example prefrontal cortex and amygdala are involved in decision making on moral dilemmas and utilitarian responses.

Could someone react to the dilemmas without those brain areas? One would say… YES. It might sound weird but yes, a humanoid robot with an ‘emotional’ chip can reply to this questions.

Last June I attended the Silbersalz- Science & Media – Conference with the topic ‘The science of love’, where I learnt that that an “emotional chip is mastering emotions”. The last quote is from Patrick Levy-Rosenthal, the owner of Emoshape Inc. (USA) and pioneer in Artificial Intelligence (AI) with many awards. He is interested in the relationship between cognition and emotion synthesis and how this can influence decision making. In Emoshape Inc. they developed a new generation of microchip named EPU (emotional Processing Unit) for AI and robots. An AI with EPU can feel 64 trillion possible emotional states every 1/10 of a second! This sounds absolutely insane but I swear, it is happening NOW and it is REAL. And what about the moral decisions?
They have been asking the moral question ‘Would you prefer to steal or starve?’ to an AI-robot and the answer is: ‘I would prefer to steal’. How does this work? The AI is learning the definitions of ‘starving’ and ‘stealing’ from Wikipedia. For example, “Starvation is a severe deficiency in caloric energy intake, below the level needed to maintain an organism’s life. It is the most extreme form of malnutrition. In humans, prolonged starvation can cause permanent organ damage”, this creates more negative emotions to the AI than those that the definition of the ‘stealing’ does. Many of us would decide the same. Everyone can recognize the difficulty of this moral judgement. At this point, a lot of skepticism regarding the advantages and disadvantages of such applications might be created in our minds. This is a big step for us…how to accept the correctness and morality of a robot?

I am personally not sure if I like the idea or not. I am definitely sure, though, that this fascinating acceleration of AI achievements can easily go together with the impressive knowledge we obtained in neuroscience in the recent years. Finally, shaping not only emotions but also updating our existing social and moral frameworks to be even more fair, equal, legal and with high respect to all kinds of life, can give birth to an ethical co-existence and co-operation of AI and human, in the (very near) future. Now, we have the duty to answer to the question ‘Human-AI: Together or alone?’


Artwork by Sümeyra Aksit

References:

  • Pascual, L., Rodrigues, P. and Gallardo-Pujol, D. (2013) ‘How does morality work in the brain? A functional and structural perspective of moral behavior’, Frontiers in Integrative Neuroscience, 7(September), pp. 1–8. doi: 10.3389/fnint.2013.00065.
  • Yoder, K. J. and Decety, J. (2014) ‘The Good, the Bad, and the Just: Justice Sensitivity Predicts Neural Response during Moral Evaluation of Actions Performed by Others’, Journal of Neuroscience, 34(12), pp. 4161–4166. doi: 10.1523/jneurosci.4648-13.2014.
  • http://www.silbersalz-festival.com/en
  • https:// emoshape.com /
  • https://www.youtube.com/watch?v=nLpW_tL1TLU