# You meet a machine that claims to be sentient. What does it take to convince you that is so?

**URL:** <https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215>\
**Category:** In My Humble Opinion\
**Created:** [January 19, 2011, 5:04pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215 "2011-01-19T17:04:05Z")\
**Posts on this page:** 20\
**Page:** 2

<div class="post-metadata">

**Author:** ![Lakai](https://avatars.discourse-cdn.com/v4/letter/l/9de0a6/32.png) [@Lakai](https://boards.straightdope.com/u/Lakai)\
**Post date:** [January 20, 2011, 2:45am UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/21 "2011-01-20T02:45:03Z")

</div>

While most humans can’t manage this, if a machine picks up on my sarcasm, I would be convinced.

---

<div class="post-metadata">

**Author:** ![AHunter3](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/ahunter3/32/368_2.png) [@AHunter3](https://boards.straightdope.com/u/AHunter3)\
**Post date:** [January 20, 2011, 4:42am UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/22 "2011-01-20T04:42:33Z")

</div>

I don’t care what it looks like. It could be a cube full of circuitry and be sentient, or could look and feel like humans and fail to be more than a machine.

I don’t have a default attitude/position of “machines can’t be sentient” so it’s hard to list any “test” one way or the other.

If it starts teasing me about the whole issue of sentience, that’s a good start.

---

<div class="post-metadata">

**Author:** ![Claverhouse](https://avatars.discourse-cdn.com/v4/letter/c/13edae/32.png) [@Claverhouse](https://boards.straightdope.com/u/Claverhouse)\
**Post date:** [January 20, 2011, 5:09am UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/23 "2011-01-20T05:09:50Z")

</div>

I would remove it’s energy supply ( AKA electricity ) and invite it to replace it.

---

<div class="post-metadata">

**Author:** ![Baal\_Houtham](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/baal_houtham/32/13604_2.png) [@Baal\_Houtham](https://boards.straightdope.com/u/Baal_Houtham)\
**Post date:** [January 20, 2011, 11:25am UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/24 "2011-01-20T11:25:55Z")

</div>

> [@Skald\_the\_Rhymer](#):
>
> One need not believe in free will to accept sentience. I mean, it’s obvious to me that **I** am self-aware. Hell, I’d call that incontrovertible.

> [@billfish678](#):
>
> Sentience, free will, and self aware are IMO not necessarily all exactly the same things.

I’m not _sure_ I have a point in tying them together, but if there’s no free will then it’s close to saying we’re robots. And if we’re “robots” are we actually self-aware and sentient?

This stuff makes my head hurt, and in a effort to stop the hurting I looked up “sentient” in the One Look Dictionary. The quick definitions were interesting –

_capable of feeling things through physical senses  
more…  
Provided by Macmillan Dictionary_

and

_Quick definitions from WordNet  
▸ adjective: endowed with feeling and unstructured consciousness  
Wordnet_

The MacMillan one is close to my initial notion of being able to feel pain (which I can’t really imagine a machine doing.) Presumably a sentient creature would also feel pleasure, but suffering is more “humanizing” than self-gratification.

---

<div class="post-metadata">

**Author:** ![MegaBee](https://avatars.discourse-cdn.com/v4/letter/m/f0a364/32.png) [@MegaBee](https://boards.straightdope.com/u/MegaBee)\
**Post date:** [January 20, 2011, 12:25pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/25 "2011-01-20T12:25:23Z")

</div>

Transforming into a truck would be a good start.

---

<div class="post-metadata">

**Author:** ![Autolycus](https://avatars.discourse-cdn.com/v4/letter/a/22d042/32.png) [@Autolycus](https://boards.straightdope.com/u/Autolycus)\
**Post date:** [January 20, 2011, 12:29pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/26 "2011-01-20T12:29:08Z")

</div>

> [@MegaBee](#):
>
> Transforming into a truck would be a good start.

How so? The series tagline is “Transformers: Robots in Disguise.” That just leaves us where we started.

---

<div class="post-metadata">

**Author:** ![MrDibble](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/mrdibble/32/114_2.png) [@MrDibble](https://boards.straightdope.com/u/MrDibble)\
**Post date:** [January 20, 2011, 12:57pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/27 "2011-01-20T12:57:26Z")

</div>

I picked “[any of them] could convince me if \<blank\>” where \<blank\> = have at least 3 sequential conversations with me that convince me it’s not just a chatbot, bearing in mind that I’m familiar with common chatbot techniques. I’d expect evidence of a degree of learning both between and within sessions, as well as some sort of empathetic response - not necessarily identical to human baseline - I’m happy with, say, doglike empathy (_catlike_ is obviously right out.)

---

<div class="post-metadata">

**Author:** ![BMalion](https://avatars.discourse-cdn.com/v4/letter/b/e274bd/32.png) [@BMalion](https://boards.straightdope.com/u/BMalion)\
**Post date:** [January 20, 2011, 2:57pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/28 "2011-01-20T14:57:40Z")

</div>

> [@MrDibble](#):
>
> I picked “[any of them] could convince me if \<blank\>” where \<blank\> = have at least 3 sequential conversations with me that convince me it’s not just a chatbot, bearing in mind that I’m familiar with common chatbot techniques. I’d expect evidence of a degree of learning both between and within sessions, as well as some sort of empathetic response - not necessarily identical to human baseline - I’m happy with, say, doglike empathy (_catlike_ is obviously right out.)

So tell me about your mother.

---

<div class="post-metadata">

**Author:** ![Balthisar](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/balthisar/32/2825_2.png) [@Balthisar](https://boards.straightdope.com/u/Balthisar)\
**Post date:** [January 20, 2011, 3:37pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/29 "2011-01-20T15:37:45Z")

</div>

I wrote a novel many, many years ago (in which I invented the iPhone and it was amazing for sporting 4096 _different_ colors!) about a sentient machine, and the problem with sentience was that it also fostered an unfortunate (for us) sense of self preservation (for it).

To convince me that it (the machine) was a person, I think I’d have to see some instinct for self preservation.

Of course, dogs are pretty smart, and have an non-intellectual instinct for self preservation, and although I love dogs and recognize that they have personality, I’m not one of those wackos that would extend personhood to dogs.

So that machine would also have to be able to communicate effectively. Dogs are extremely limited in their ability to communicate in a mutual manner with humans, and so machine communication would have to be much more expressive.

Creativity may be the key. I’m not sure that dogs are creative. They’re observant and able to learn, but I’m not certain that they’re creative or are able to innovate on their own.

Heck, suppose we’re all a computer simulation? Are we persons according to the programmer?

---

<div class="post-metadata">

**Author:** ![outlierrn](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/outlierrn/32/3005_2.png) [@outlierrn](https://boards.straightdope.com/u/outlierrn)\
**Post date:** [January 20, 2011, 6:32pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/30 "2011-01-20T18:32:16Z")

</div>

> [@Balthisar](#):
>
> and have an non-intellectual instinct for self preservation

What’s your basis for this statement?

---

<div class="post-metadata">

**Author:** ![Peremensoe](https://avatars.discourse-cdn.com/v4/letter/p/ce7236/32.png) [@Peremensoe](https://boards.straightdope.com/u/Peremensoe)\
**Post date:** [January 20, 2011, 6:40pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/31 "2011-01-20T18:40:53Z")

</div>

I’m baffled by the choices of different types of “unit.” Who thinks _shape_ is relevant here?

---

<div class="post-metadata">

**Author:** ![Harmonious\_Discord](https://avatars.discourse-cdn.com/v4/letter/h/74df32/32.png) [@Harmonious\_Discord](https://boards.straightdope.com/u/Harmonious_Discord)\
**Post date:** [January 20, 2011, 6:55pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/32 "2011-01-20T18:55:26Z")

</div>

The machine has to be able to figure out how to convince me or it isn’t sentient.

---

<div class="post-metadata">

**Author:** ![Balthisar](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/balthisar/32/2825_2.png) [@Balthisar](https://boards.straightdope.com/u/Balthisar)\
**Post date:** [January 20, 2011, 7:32pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/33 "2011-01-20T19:32:15Z")

</div>

> [@outlierrn](#):
>
> What’s your basis for this statement?

Because dogs won’t intentionally commit suicide. They may not know that not all cars will stop from running them over and die, but they don’t get killed intentionally. They don’t randomly fall off cliffs. They don’t meekly submit when attacked.

---

<div class="post-metadata">

**Author:** ![cochrane](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/cochrane/32/10441_2.png) [@cochrane](https://boards.straightdope.com/u/cochrane)\
**Post date:** [January 20, 2011, 7:45pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/34 "2011-01-20T19:45:38Z")

</div>

Are you allowing non-physical representations of sentient beings such as holograms? I propose that self-aware holograms. such as the EMH from Voyager, Vic Fontaine from DS9, and Moriarty from those STTNG episodes where Data plays Sherlock Holmes in the holodeck could pass themselves off as sentient. Voyager’s Doctor can even leave the Sickbay, thanks to his portable emitter and wander around freely as he likes. I could easily accept any of these entities as sentient, particularly the EMH.

---

<div class="post-metadata">

**Author:** ![Walmarticus](https://avatars.discourse-cdn.com/v4/letter/w/f9ae1b/32.png) [@Walmarticus](https://boards.straightdope.com/u/Walmarticus)\
**Post date:** [January 20, 2011, 7:53pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/35 "2011-01-20T19:53:00Z")

</div>

Mike almost certainly displayed that he was capable of feelings by having a personal interest in seeing the rebellion succeed. His character of Adam Salene was as vested in the physical reality of the colony as any true man. He displayed original, creative thought for his own rational ends, as well as those of his friends. He sacrificed himself for the greater good. He was a loyal patriot and a true friend.

_sniff_. I can’t go on.

---

<div class="post-metadata">

**Author:** ![Apollyon](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/apollyon/32/4238_2.png) [@Apollyon](https://boards.straightdope.com/u/Apollyon)\
**Post date:** [January 20, 2011, 10:20pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/36 "2011-01-20T22:20:40Z")

</div>

> [@Claverhouse](#):
>
> I would remove it’s energy supply ( AKA electricity ) and invite it to replace it.

Would you be OK with the machine asking you to submit to the same test?

I’m just reminded of the golem Dorfl in **Pratchett’s** _Feet of Clay_ when confronted by the priests who are disputing that he is alive. Dorfl agrees that were they to grind him up until all that remained was ceramic dust, and then sift through that dust they would not find a single iota of life… and was willing to submit to this test… provided they also would submit to it.

---

<div class="post-metadata">

**Author:** ![Daerlyn](https://avatars.discourse-cdn.com/v4/letter/d/898d66/32.png) [@Daerlyn](https://boards.straightdope.com/u/Daerlyn)\
**Post date:** [January 20, 2011, 10:29pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/37 "2011-01-20T22:29:26Z")

</div>

> [@Apollyon](#):
>
> Would you be OK with the machine asking you to submit to the same test?
> 
> I’m just reminded of the golem Dorfl in **Pratchett’s** _Feet of Clay_ when confronted by the priests who are disputing that he is alive. Dorfl agrees that were they to grind him up until all that remained was ceramic dust, and then sift through that dust they would not find a single iota of life… and was willing to submit to this test… provided they also would submit to it.

“Is it me, or are we on shaky theological ground here?” 😃  
The self-preservation concept is a good one. I’d expand that into intellectual/emotional self-preservation/indulgence as well. Can the machine recognise certain activities and experiences as enjoyable or fun, and others as unpleasant, boring, or painful? If it’s able to form and express preferences like that in a plausible way, I’d be willing to ascribe sentience/sapience/whatever until proven otherwise.

---

<div class="post-metadata">

**Author:** ![Fear\_Itself](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/fear_itself/32/19637_2.png) [@Fear\_Itself](https://boards.straightdope.com/u/Fear_Itself)\
**Post date:** [January 20, 2011, 11:23pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/38 "2011-01-20T23:23:08Z")

</div>

Tell it, “Everything I say is a lie.” If its head explodes in a cloud of smoke and sparks, it was a robot.

---

<div class="post-metadata">

**Author:** ![Superfluous\_Parentheses](https://avatars.discourse-cdn.com/v4/letter/s/8edcca/32.png) [@Superfluous\_Parentheses](https://boards.straightdope.com/u/Superfluous_Parentheses)\
**Post date:** [January 20, 2011, 11:27pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/39 "2011-01-20T23:27:13Z")

</div>

“Simple” Turing test; if I can have a meaningful series of conversations about a wide range of topics with any one of the proposed subjects, and I cannot distinguish their part of the conversations from - say - a human “speaking through them” remotely, they are sentient. And I think the Turing test is somewhat stricter than it needs to be in this case, since the test assumes a human-equivalent level of intelligence and experience, but it certainly would be sufficient.

---

<div class="post-metadata">

**Author:** ![Superfluous\_Parentheses](https://avatars.discourse-cdn.com/v4/letter/s/8edcca/32.png) [@Superfluous\_Parentheses](https://boards.straightdope.com/u/Superfluous_Parentheses)\
**Post date:** [January 20, 2011, 11:40pm UTC](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215/40 "2011-01-20T23:40:52Z")

</div>

> [@TriPolar](#):
>
> If I know the physical nature of this entity, I know it’s not human, and not as much a person as a human.

So? The question is not if it’s human.

> [@](#):
>
> If such a machine had been manufactured, and operated in such a way as to simulate all human emotional and rational processes, and nothing else,

How would you know? I’m quite certain that if we ever build a machine complicated enough to do all that, we wouldn’t be able to say conclusively that it’s limited to just “simulation” of whatever we “programmed” into it.

> [@](#):
>
> […]it has no reason to appear to humans as if it is anything but a machine, and nobody should have worry about hurting its feelings or otherwise cause it to suffer in it’s own perception. Humans don’t have a choice in their condition, but machines would not be so constrained.

I don’t follow you. Bacteria are more like machines than we are, but they also seem to be more constrained than us, at least as far as intention is concerned.

[Previous page](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215.md?page=1)

[Next page](https://boards.straightdope.com/t/you-meet-a-machine-that-claims-to-be-sentient-what-does-it-take-to-convince-you-that-is-so/568215.md?page=3)
