Editing Sentience (section)

== Digital sentience ==
{{Main articles|Artificial consciousness}}

Digital sentience (or artificial sentience) means the sentience of [[Artificial intelligence|artificial intelligences]]. The question of whether artificial intelligences can be sentient is controversial.<ref name=":0">{{Cite news |last=Jackson |first=Lauren |date=2023-04-12 |title=What if A.I. Sentience Is a Question of Degree? |language=en-US |work=The New York Times |url=https://www.nytimes.com/2023/04/12/world/artificial-intelligence-nick-bostrom.html |access-date=2023-06-23 |issn=0362-4331}}</ref> 

The AI research community does not consider sentience (that is, the "ability to feel sensations") as an important research goal, unless it can be shown that consciously "feeling" a sensation can make a machine more intelligent than just receiving input from sensors and processing it as information. [[Stuart J. Russell|Stuart Russell]] and [[Peter Norvig]] wrote in 2021: "We are interested in programs that behave intelligently. Individual aspects of consciousness—awareness, self-awareness, attention—can be programmed and can be part of an intelligent machine. The additional project making a machine conscious in exactly the way humans are is not one that we are equipped to take on."{{sfn|Russell|Norvig|2021|p=986}} 
Indeed, leading AI textbooks do not mention "sentience" at all.<ref>Leading AI textbooks in 2023:
* {{Cite book
|first1=Stuart J.|last1=Russell|author1-link=Stuart J. Russell
|first2=Peter|last2=Norvig|author2-link=Peter Norvig
|title=[[Artificial Intelligence: A Modern Approach]]
|year=2021
|edition=4th
|isbn=9780134610993
|lccn=20190474
|publisher=Pearson|location=Hoboken
}}
* {{Cite book
| first1 = Elaine | last1 = Rich | author1-link = Elaine Rich
| first2 = Kevin  | last2 = Knight
| first3 = Shivashankar B | last3 = Nair
| title = Artificial Intelligence
| date = 1 January 2010
| isbn = 9780070087705
| language = en
| publisher = Tata McGraw Hill India | location = New Delhi
| edition = 3rd
}}
</ref>

Digital sentience is of considerable interest to the [[philosophy of mind]]. [[Functionalism (philosophy of mind)|Functionalist]] philosophers consider that sentience is about "causal roles" played by mental states, which involve information processing. In this view, the physical substrate of this information processing does not need to be biological, so there is no theoretical barrier to the possibility of sentient machines.<ref>{{Cite journal |last1=Manzotti |first1=Riccardo |last2=Chella |first2=Antonio |date=2018 |title=Good Old-Fashioned Artificial Consciousness and the Intermediate Level Fallacy |journal=Frontiers in Robotics and AI |volume=5 |page=39 |doi=10.3389/frobt.2018.00039 |pmid=33500925 |pmc=7805708 |issn=2296-9144 |doi-access=free }}</ref> According to [[type physicalism]] however, the physical constitution is important; and depending on the types of physical systems required for sentience, it may or may not be possible for certain types of machines (such as electronic computing devices) to be sentient.<ref>{{Cite journal |last=Searle |first=John R. |date=1980 |title=Minds, brains, and programs |url=http://members.aol.com/NeoNoetics/MindsBrainsPrograms.html |journal=Behavioral and Brain Sciences |language=en |volume=3 |issue=3 |pages=417–424 |doi=10.1017/S0140525X00005756 |s2cid=55303721 |issn=1469-1825|archive-url=https://web.archive.org/web/20071210043312/http://members.aol.com/NeoNoetics/MindsBrainsPrograms.html |archive-date=2007-12-10 }}</ref>

The discussion on the topic of alleged sentience of artificial intelligence has been reignited in 2022 by the claims made about [[Google]]'s [[LaMDA]] ([[Language model|Language Model]] for Dialogue Applications) artificial intelligence system that it is "sentient" and had a "[[soul]]".<ref>{{Cite web |last=Brandon Specktor published |date=2022-06-13 |title=Google AI 'is sentient,' software engineer claims before being suspended |url=https://www.livescience.com/google-sentient-ai-lamda-lemoine |access-date=2022-06-14 |website=livescience.com |language=en}}</ref> LaMDA is an [[Artificial Intelligence System|artificial intelligence system]] that creates [[chatbot]]s—AI robots designed to communicate with humans—by gathering vast amounts of text from the internet and using [[Algorithm|algorithms]] to respond to queries in the most fluid and natural way possible. The transcripts of conversations between scientists and LaMDA reveal that the AI system excels at this, providing answers to challenging topics about the nature of [[Emotion|emotions]], generating [[Aesop's Fables|Aesop]]-style fables on cue, and even describing its alleged fears.<ref>{{Cite web |last=Lemoine |first=Blake |date=2022-06-11 |title=Is LaMDA Sentient? — an Interview |url=https://cajundiscordian.medium.com/is-lamda-sentient-an-interview-ea64d916d917 |access-date=2022-06-14 |website=Medium |language=en}}</ref>

[[Nick Bostrom]] considers that while LaMDA is probably not sentient, being very sure of it would require understanding how consciousness works, having access to unpublished information about LaMDA's architecture, and finding how to apply the philosophical theory to the machine.<ref>{{Cite web |last=Leith |first=Sam |date=2022-07-07 |title=Nick Bostrom: How can we be certain a machine isn't conscious? |url=https://www.spectator.co.uk/article/nick-bostrom-how-can-we-be-certain-a-machine-isnt-conscious/ |access-date=2023-06-23 |website=The Spectator |language=en-US}}</ref> He also said about LLMs that "it's not doing them justice to say they're simply regurgitating text", noting that they "exhibit glimpses of creativity, insight and understanding that are quite impressive and may show the rudiments of reasoning". He thinks that "sentience is a matter of degree".<ref name=":0" />

In 2022, philosopher [[David Chalmers]] made a speech on whether [[large language model|large language models]] (LLMs) can be conscious, encouraging more research on the subject. He suggested that current LLMs were probably not conscious, but that the limitations are temporary and that future systems could be serious candidates for consciousness.<ref>{{Cite news |last=Chalmers |first=David |date=August 9, 2023 |title=Could a Large Language Model Be Conscious? |url=https://www.bostonreview.net/articles/could-a-large-language-model-be-conscious/ |work=Boston Review}}</ref>

According to [[Jonathan Birch (philosopher)|Jonathan Birch]], "measures to regulate the development of sentient AI should run ahead of what would be proportionate to the risks posed by current technology, considering also the risks posed by credible future trajectories." He is concerned that AI sentience would be particularly easy to deny, and that if achieved, humans might nevertheless continue to treat AI systems as mere tools. He notes that the linguistic behaviour of LLMs is not a reliable way to assess whether they are sentient. He suggests to apply theories of consciousness, such as the [[global workspace theory]], to the algorithms implicitly learned by LLMs, but noted that this technique requires advances in [[AI interpretability]] to understand what happens inside. He also mentions some other pathways that may lead to AI sentience, such as the [[brain emulation]] of sentient animals.<ref>{{Cite book |last=Birch |first=Jonathan |url=https://academic.oup.com/book/57949 |title=The Edge of Sentience: Risk and Precaution in Humans, Other Animals, and AI |date=19 July 2024 |publisher=Oxford University Press |chapter=Part V: preparing for artificial sentience|doi=10.1093/9780191966729.001.0001 |isbn=978-0-19-196672-9 }}</ref>