AI for Death and Dying

Maggi Savin-Baden

• Publisher: CRC Press
• Year: 2021
• ISBN:
  9781000435726
• Edition: 1
• Topic: Computer Science
• Subtopic: Artificial Intelligence (AI) & Semantics
• Index: Computer Science

What is artificial intelligence (AI)? How does AI affect death matters and the digital beyond? How are death and dying handled in our digital age?

AI for Dying and Death covers a broad range of literature, research and challenges around this topic. It explores ethical memorisation, digital legacies and bereavement, post death avatars and AI and the digital beyond. It also analyzes religious perspectives on AI for death and dying, and planning for death in a digital age.

1
What Is Artificial Intelligence?

Introduction

This chapter begins by setting artificial intelligence in the context of current conceptions as it is most notable that the idea of artificial intelligence tends to prompt extreme views as well the creation of false hope and concerns. It then examines the various forms of artificial intelligence by defining five different types as the basis for understanding how in later chapters artificial intelligence is seen and used in relation to death and dying. The final section of the chapter suggests that there are a number of challenges raised by the diverse types of artificial intelligence and concludes that automated decision-making is probably more of a concern in the current climate that artificial intelligence.

Portrayal and Perceptions of Artificial Intelligence

Artificial intelligence (AI) is often thought of as robots or thinking machines. In marketing, it is seen as a programme or algorithm often based on machine-learning principles, yet the complexity and diversity of AI is much broader than this. Consequently, AI can be grouped broadly into three main types: first, deep AI which involves creating virtual humans and digital immortals with sentience, the ability to feel, perceive, or experience; second, pragmatic AI where machines are used for tasks such as mining large data sets, or automating complex or risky tasks; and third, marketing AI where companies use algorithms to anticipate the customer’s next move and improve the customer journey. Often when artificial intelligence (AI) is spoken about in the 21st century, the perspectives that many people have gained are from information presented by media and film. Examples of these are provided in Table 1.1. Furthermore, Recchia (2020) analysed a data set of over 100,000 film subtitles and identified control (or loss of it) as a recurring motif in films about artificial intelligence, which tends to reflect broad public opinion about AI in general (Burden & Savin-Baden, 2019).

Table 1.1 Artificial Intelligence in Films and Media

Character	Film/Media (Date Order)	Notion of AI	Difficulty
Hal	2001: A Space Odyssey (Kubrick & Clarke, 1968)	Hal controls a spaceship, and prioritizes the mission over the crew	Setting goals for an AI that do not have unintended consequences
Replicants	Blade Runner (Deeley & Scott, 1982)	Androids that go rogue, and how to spot them	It is not clear if a replicant knows that it is a replicant.
Data	Star Trek (Berman, 1995)	Sentient android as crew mate	Does it have real emotions? Is it too artificial?
Nelson	The Red Men (De Abaitua, 2009)	Creating a virtual town of people in order to model human behaviour.	What happens if the computer starts modelling the wrong things – like WW1?
Zoe	Caprica (Aubuchon & Moore, 2010)	Making a copy of oneself – and how that copy feels when the creator dies prematurely. Building the copy from digital traces. The importance of a virtual world as ‘home’.	The need for a body (Cylons). Is there enough information in the digital traces?
Ash	Be Right Back (Brooker & Harris, 2013).	The character Ash is built on the basis of his social media profile	It is implied that AI beings can/will have highly advanced emotional responses and empathy.
Ava	Ex Machina (Macdonald & Garland, 2015)	That it is possible to create a true copy of human intelligence	The question of whether a female robot is considered to have human-like intelligence is left to the judgement of a non-expert man selected based on the fact that he will be sexually attracted to her.
Will	Transcendence (Kosove et al., (2014)	Will is shot, and his wife uploads his consciousness into a quantum computer	It presents the idea of a superhuman AI being who transcends the laws of physics
Leila	Kiss Me First (Moggach, 2014)	That massively multiplayer online role-playing games can influence real-world characters so that the real and the fictional activities collide	The person who is killed in real life is still living in the here and now of the game, acting in similar ways as their real-life persona
Ultron	Avengers: Age of Ultron Feige and Whedon (2015)	AI can be used to create a complete copy of a human, but more intelligent.	Ultron is sentient rather than being created as digital immortal personas of other people
Daughter	C4 Humans (Vincent & Brackley, 2015)	A scientist tries to recreate a copy of her dead daughter	There does not seem to be enough information or sophistication to keep it stable
Synths	C4 Humans (Vincent & Brackley, 2015)	Creating worker androids, but then endowing some of them with sentience	The right for the sentient synths to procreate. Whether it is ethical to uplift non-sentient synths.
Girlfriend	Blade Runner 2049 (Kosove & Villeneuve 2017)	AI as a virtual friend/girlfriend.	Limited agency when not needed. Overlaying with a physical sex worker.
Rachel	The After wife (Hunter, 2018)	Rachel (the AI) is created by human Rachel before she dies and without any knowledge or consent from her family	This book raises ethical concerns about not only the rights of the robot, but also about whether family should be consulted pre-death about the creation of a copy
Imagos	Memory of Empire (Martine, 2019)	Creating a copy of your predecessor in a post that exists alongside you in your body as a computer, plugged into your central nervous system	What happens if it overloads or cooks the brain? It is not clear what agency it will be given over the physical body
Daughter	Devs (Garland, 2020)	A dot com billionaire tries to recreate his dead daughter	More by modelling and an assumption of predetermination.

Recently, there have been considerable improvements to AI such as better text-to-speech, improved speech recognition and high-quality avatars. Other stances towards AI tend to be much more technical and are based on assumptions about the possibilities for uploading the brain or alternatively using algorithms to predict behaviour and shopping patterns. Frude and Jandric (2015) argue that the real shift in AI will occur when interaction between humans and AI is cumulative. This will result in long-term relationships with the person trusting the machine and in turn the machine adapting to appreciate the person’s preferences. However, it is important to note that media and press reports still tend to imply that AI is more developed than it actually is, and it is important to be able to distinguish between clever algorithms, some form of well-developed virtual human and the possibility of creating a virtual persona with sentience some time in the next 40 years.

Types of Artificial Intelligence

The early days of research into AI focused on creating computer programmes that could ‘understand’ and manipulate symbols in order to reason; this was termed symbolic AI. Other forms are embodied AI and enactive AI. Embodied AI is where software is imported into a physical (robotic) body and exploring how that body fits into real-world environments and thus is based on the idea of embodied cognition – the view that intelligence is as much a part of the body as the brain. Enactive AI is seen as a form of AI that refines and complements embodied AI because it has a focus on the self-identity of the robot or virtual human, so that meaning is created by the interaction of the robot with the environment (Froese & Ziemke, 2009). There is also often confusion between AI and augmented intelligence. Augmented intelligence describes either using a computer that is external to the human to help the human with mental tasks or inking computing capability directly into the brain in order to enhance the brain’s processing power or memory. The focus here is broadly on the use of AI to help humans with a task, but it also extends to areas such as embodiment and post-digital humans, as these areas have direct impact on the relationship between AI and death and dying. However, whilst AI can be identified through terms such as deep, pragmatic, marketing, embodied or enactive AI, it can also be categorized in terms of its impact and effect. This section positions AI as something that includes a range of concepts from algorithms to posthumanism.

Algorithms and Machine Learning

Simple algorithms are used to create a series of steps to undertake a task which in general is straightforward and predictable. These kinds of algorithms often appear sophisticated but in general are not so and cannot really be classed as AI. An example would be Facebook algorithms designed to advertise to individuals, based on their browsing and shopping history. Complex algorithms are often the...

Table of contents

• Cover
• Half Title
• Series Page
• Title Page
• Copyright Page
• Dedication
• Table of Contents
• List of Figures
• List of Tables
• Permissions
• Acknowledgements
• Author
• Introduction
• 1 What Is Artificial Intelligence?
• 2 Death and Dying in the Digital Age
• 3 Ethical Memorialization
• 4 Digital Legacy Management
• 5 Religious Perspectives
• 6 Digital Bereavement
• 7 Digital Afterlife, Digital Immortal Creation and Artificial Intelligence
• 8 Artificial Intelligence and the Digital Beyond
• References
• Index