The Turing test: Can a computer pass for a human? - Alex Gendler

1,681,348 views ใƒป 2016-04-25

TED-Ed


ืื ื ืœื—ืฅ ืคืขืžื™ื™ื ืขืœ ื”ื›ืชื•ื‘ื™ื•ืช ื‘ืื ื’ืœื™ืช ืœืžื˜ื” ื›ื“ื™ ืœื”ืคืขื™ืœ ืืช ื”ืกืจื˜ื•ืŸ.

ืชืจื’ื•ื: Ido Dekkers ืขืจื™ื›ื”: Roni Ravia
00:06
What is consciousness?
0
6999
2035
ืžื” ื”ื™ื ืชื•ื“ืขื”?
00:09
Can an artificial machine really think?
1
9034
2794
ื”ืื ืžื›ื•ื ื” ืžืœืื›ื•ืชื™ืช ื™ื›ื•ืœื” ื‘ืืžืช ืœื—ืฉื•ื‘?
00:11
Does the mind just consist of neurons in the brain,
2
11828
3286
ื”ืื ืชื‘ื•ื ื” ืžื•ืจื›ื‘ืช ืจืง ืžืชืื™ ืขืฆื‘ ื‘ืžื•ื—
00:15
or is there some intangible spark at its core?
3
15114
3643
ืื• ืฉืžื ื”ื™ื ื—ื•ื‘ื” ื‘ืœื™ื‘ื” ื ื™ืฆื•ืฅ ื‘ืœืชื™ ื ืชืคืก?
00:18
For many, these have been vital considerations
4
18757
2568
ืขื‘ื•ืจ ืจื‘ื™ื, ืืœื” ื”ื™ื• ืฉื™ืงื•ืœื™ื ื—ืฉื•ื‘ื™ื
00:21
for the future of artificial intelligence.
5
21325
3027
ืœืขืชื™ื“ื” ืฉืœ ื”ื‘ื™ื ื” ื”ืžืœืื›ื•ืชื™ืช.
00:24
But British computer scientist Alan Turing decided to disregard all these questions
6
24352
5187
ืื‘ืœ ืžื“ืขืŸ ื”ืžื—ืฉื‘ื™ื ื”ื‘ืจื™ื˜ื™ ืืœืŸ ื˜ื•ืจื™ื ื’ ื”ื—ืœื™ื˜ ืœื”ืชืขืœื ืžื”ืฉืืœื•ืช ื”ืืœื•
00:29
in favor of a much simpler one:
7
29539
2311
ืœื˜ื•ื‘ืช ืฉืืœื” ื”ืจื‘ื” ื™ื•ืชืจ ืคืฉื•ื˜ื”:
00:31
can a computer talk like a human?
8
31850
3419
ื”ืื ืžื—ืฉื‘ ื™ื›ื•ืœ ืœื“ื‘ืจ ื›ืžื• ื‘ืŸ ืื ื•ืฉ?
00:35
This question led to an idea for measuring aritificial intelligence
9
35269
3996
ืฉืืœื” ื–ื• ื’ืจืžื” ืœื• ืœื”ืžืฆื™ื ืฉื™ื˜ื” ืœืžื“ื™ื“ืช ื‘ื™ื ื” ืžืœืื›ื•ืชื™ืช
00:39
that would famously come to be known as the Turing test.
10
39265
4051
ืฉื”ืชืคืจืกืžื” ืœืื—ืจ ืžื›ืŸ ื‘ืฉื ืžื‘ื—ืŸ ื˜ื•ืจื™ื ื’.
00:43
In the 1950 paper, "Computing Machinery and Intelligence,"
11
43316
4025
ื‘ืžืืžืจ ืž-1950, "ืžื›ื•ื ื•ืช ืžื—ืฉื‘ ื•ื‘ื™ื ื”",
00:47
Turing proposed the following game.
12
47341
2473
ื˜ื•ืจื™ื ื’ ื”ืฆื™ืข ืืช ื”ืžืฉื—ืง ื”ื‘ื.
00:49
A human judge has a text conversation with unseen players
13
49814
4092
ืฉื•ืคื˜ ืื ื•ืฉื™ ืžื ื”ืœ ืฉื™ื—ื” ื‘ื›ืชื‘ ืขื ืฉื—ืงื ื™ื ื‘ืœืชื™ ื ืจืื™ื
00:53
and evaluates their responses.
14
53906
2476
ื•ืžืขืจื™ืš ืืช ืชื’ื•ื‘ื•ืชื™ื”ื.
00:56
To pass the test, a computer must be able to replace one of the players
15
56382
4027
ื›ื“ื™ ืœืขื‘ื•ืจ ืืช ื”ืžื‘ื—ืŸ, ืžื—ืฉื‘ ื—ื™ื™ื‘ ืœื”ื™ื•ืช ืžืกื•ื’ืœ ืœื”ื—ืœื™ืฃ ืืช ืื—ื“ ื”ืฉื—ืงื ื™ื
01:00
without substantially changing the results.
16
60409
3448
ืžื‘ืœื™ ืœืฉื ื•ืช ืืช ื”ืชื•ืฆืื•ืช ื‘ืฆื•ืจื” ืžืฉืžืขื•ืชื™ืช.
01:03
In other words, a computer would be considered intelligent
17
63857
2959
ื‘ืžื™ืœื™ื ืื—ืจื•ืช, ืžื—ืฉื‘ ื™ื—ืฉื‘ ืื™ื ื˜ื™ืœื™ื’ื ื˜ื™
01:06
if its conversation couldn't be easily distinguished from a human's.
18
66816
5697
ืื ืœื ื™ื”ื™ื” ื ื™ืชืŸ ืœื”ื‘ื“ื™ืœ ืชื•ืš ื›ื“ื™ ืฉื™ื—ื” ื‘ื™ื ื• ืœื‘ื™ืŸ ื‘ืŸ ืื ื•ืฉ.
01:12
Turing predicted that by the year 2000,
19
72513
2154
ื˜ื•ืจื™ื ื’ ื—ื–ื” ืฉืขื“ ืฉื ืช 2000,
01:14
machines with 100 megabytes of memory would be able to easily pass his test.
20
74667
5955
ืžื›ื•ื ื•ืช ืขื 100 ืžื’ื‘ื™ื™ื˜ ืฉืœ ื–ื™ื›ืจื•ืŸ ืชื”ื™ื™ื ื” ืžืกื•ื’ืœื•ืช ืœืขื‘ื•ืจ ืืช ื”ืžื‘ื—ืŸ ื‘ืงืœื•ืช.
01:20
But he may have jumped the gun.
21
80622
2164
ืื‘ืœ ืื•ืœื™ ื”ื•ื ื”ืงื“ื™ื ืืช ื”ืžืื•ื—ืจ ืขื ื”ืฆื”ืจื” ื–ื•.
01:22
Even though today's computers have far more memory than that,
22
82786
3054
ืืคื™ืœื• ืฉืœืžื—ืฉื‘ื™ื ืฉืœ ื”ื™ื•ื ื™ืฉ ื”ืจื‘ื” ื™ื•ืชืจ ื–ื™ื›ืจื•ืŸ,
01:25
few have succeeded,
23
85840
1803
ืžืขื˜ื™ื ืฆืœื—ื• ืืช ื”ืžื‘ื—ืŸ,
01:27
and those that have done well
24
87643
1795
ื•ืืœื” ืฉืขืฉื• ื–ืืช,
01:29
focused more on finding clever ways to fool judges
25
89438
3770
ื”ืชืžืงื“ื• ื™ื•ืชืจ ื‘ืžืฆื™ืืช ื“ืจื›ื™ื ืžืชื•ื—ื›ืžื•ืช ืœื”ื•ื ื•ืช ืฉื•ืคื˜ื™ื
01:33
than using overwhelming computing power.
26
93208
2966
ื•ืคื—ื•ืช ื‘ืฉื™ืžื•ืฉ ื‘ื›ื•ื— ืžื—ืฉื•ื‘ ืขืฆื•ื.
01:36
Though it was never subjected to a real test,
27
96174
2922
ืœืžืจื•ืช ืฉื”ื™ื ืžืขื•ืœื ืœื ืขืžื“ื” ื‘ืคื ื™ ืžื‘ื—ืŸ ืืžื™ืชื™,
01:39
the first program with some claim to success was called ELIZA.
28
99096
4769
ื”ืชื•ื›ื ื” ื”ืจืืฉื•ื ื” ืฉื ื—ืœื” ื”ืฆืœื—ื” ื ืงืจืื” "ืืœื™ื™ื–ื”".
01:43
With only a fairly short and simple script,
29
103865
2547
ืขื ืžืœืœ ืงืฆืจ ื•ืคืฉื•ื˜,
01:46
it managed to mislead many people by mimicking a psychologist,
30
106412
4022
ื”ื™ื ื”ืฆืœื™ื—ื” ืœืฉื˜ื•ืช ื‘ื”ืจื‘ื” ืื ืฉื™ื ืขืœ ื™ื“ื™ ื›ืš ืฉื”ืชื—ืงืชื” ืœืคืกื™ื›ื•ืœื•ื’ื™ืช,
01:50
encouraging them to talk more
31
110434
1961
ืขื•ื“ื“ื” ืื•ืชื ืœื“ื‘ืจ ื”ืจื‘ื”,
01:52
and reflecting their own questions back at them.
32
112395
3510
ื•ื’ืจืžื” ืœื”ื ืœื”ืชืขืžืง ื‘ืฉืืœื•ืช ืฉืœ ืขืฆืžื.
01:55
Another early script PARRY took the opposite approach
33
115905
3507
ืžืœืœ ื ื•ืกืฃ ื‘ืฉื "ืคืจื™" ื ืงื˜ ื‘ื’ื™ืฉื” ืžื ื•ื’ื“ืช
01:59
by imitating a paranoid schizophrenic
34
119412
2801
ืขืœ ื™ื“ื™ ื—ื™ืงื•ื™ ืกื›ื™ื–ื•ืคืจืŸ ืคืจื ื•ืื™ื“ื™
02:02
who kept steering the conversation back to his own preprogrammed obsessions.
35
122213
5726
ืฉื›ืœ ื”ื–ืžืŸ ืžื›ื•ื•ืŸ ืืช ืžื”ืœืš ื”ืฉื™ื—ื” ื—ื–ืจื” ืœืื•ื‘ืกืกื™ื•ืช ืžืชื•ื›ื ืชื•ืช ืžืจืืฉ.
02:07
Their success in fooling people highlighted one weakness of the test.
36
127939
4864
ื”ืขื•ื‘ื“ื” ืฉื”ืฆืœื™ื—ื• ืœืฉื˜ื•ืช ื‘ืื ืฉื™ื ื”ื‘ืœื™ื˜ื” ืืช ื—ื•ืœืฉืชื• ืฉืœ ื”ืžื‘ื—ืŸ.
02:12
Humans regularly attribute intelligence to a whole range of things
37
132803
4630
ืื ืฉื™ื ื ื•ื˜ื™ื ืœืฉื™ื™ืš ืื™ื ื˜ื™ืœื™ื’ื ืฆื™ื” ืœื˜ื•ื•ื— ืจื—ื‘ ืฉืœ ืขืฆืžื™ื
02:17
that are not actually intelligent.
38
137433
3643
ืฉืื™ื ื ืื™ื ื˜ื™ืœื™ื’ื ื˜ื™ื™ื ื›ืœืœ ื•ื›ืœืœ.
02:21
Nonetheless, annual competitions like the Loebner Prize,
39
141076
3213
ืขื ื–ืืช, ืชื—ืจื•ื™ื•ืช ืฉื ืชื™ื•ืช ื›ืžื• ืคืจืก ืœื•ื‘ื ืจ,
02:24
have made the test more formal
40
144289
1720
ื”ืคื›ื• ืืช ื”ืžื‘ื—ืŸ ืœืจืฉืžื™ ื™ื•ืชืจ
02:26
with judges knowing ahead of time
41
146009
2146
ืขื ืฉื•ืคื˜ื™ื ืฉื™ื•ื“ืขื™ื ืžืจืืฉ
02:28
that some of their conversation partners are machines.
42
148155
3864
ืฉื›ืžื” ืžืฉื•ืชืคื™ื”ื ืœืฉื™ื—ื” ื”ื ืžื›ื•ื ื•ืช.
02:32
But while the quality has improved,
43
152019
1901
ืื‘ืœ ื‘ืขื•ื“ ื”ืื™ื›ื•ืช ื”ืฉืชืคืจื”,
02:33
many chatbot programmers have used similar strategies to ELIZA and PARRY.
44
153920
5176
ืจื‘ื™ื ืžืžืชื›ื ืชื™ ื”ืจื•ื‘ื•ื˜ื™ื ื”ืฉืชืžืฉื• ื‘ืืกื˜ืจื˜ื’ื™ื•ืช ื“ื•ืžื•ืช ืœ"ืืœื™ื™ื–ื”" ื•"ืคืจื™".
02:39
1997's winner Catherine
45
159096
2036
"ืงืชืจื™ืŸ" ื”ืชื•ื›ื ื” ืฉื–ื›ืชื” ื‘-1997
02:41
could carry on amazingly focused and intelligent conversation,
46
161132
4081
ื™ื›ืœื” ืœื ื”ืœ ืฉื™ื—ื” ืžืžื•ืงื“ืช ื•ืื™ื ื˜ื™ืœื™ื’ื ื˜ื™ืช ืœื”ืคืœื™ื,
02:45
but mostly if the judge wanted to talk about Bill Clinton.
47
165213
3962
ืื‘ืœ ื‘ืขื™ืงืจ ืื ื”ืฉื•ืคื˜ ืจืฆื” ืœื“ื‘ืจ ืขืœ ื‘ื™ืœ ืงืœื™ื ื˜ื•ืŸ.
02:49
And the more recent winner Eugene Goostman
48
169175
2638
ื•ืœืื—ื“ ื”ื–ื•ื›ื™ื ื”ืื—ืจื•ื ื™ื, ื™ื•ื’'ื™ืŸ ื’ื•ืกื˜ืžืŸ
02:51
was given the persona of a 13-year-old Ukrainian boy,
49
171813
3736
ื ื™ืชื ื” ืื™ืฉื™ื•ืช ืฉืœ ื™ืœื“ ืื•ืงืจืื™ื ื™ ื‘ืŸ 13,
02:55
so judges interpreted its nonsequiturs and awkward grammar
50
175549
4021
ื›ื“ื™ ืฉื”ืฉื•ืคื˜ื™ื ื™ืคืจืฉื• ืืช ื›ืฉืœื™ื• ื”ืœื•ื’ื™ื™ื ื•ื“ืงื“ื•ืงื• ื”ืžืฉื•ื ื”
02:59
as language and culture barriers.
51
179570
3346
ื›ืžื—ืกื•ืžื™ ืฉืคื” ื•ืชืจื‘ื•ืช.
03:02
Meanwhile, other programs like Cleverbot have taken a different approach
52
182916
4219
ื‘ื™ื ืชื™ื™ื, ืžืชื›ื ืชื™ื ืื—ืจื™ื ื›ืžื• ืงืœื‘ืจื‘ื•ื˜ ื ืงื˜ื• ื‘ื’ื™ืฉื” ืฉื•ื ื”
03:07
by statistically analyzing huge databases of real conversations
53
187135
4605
ืขื ื ื™ืชื•ื— ืกื˜ื˜ื™ืกื˜ื™ ืฉืœ ืžืื’ืจื™ ืžื™ื“ืข ืขืฆื•ืžื™ื ืฉืœ ืฉื™ื—ื•ืช ืืžื™ืชื™ื•ืช ืฉืงืจื•
03:11
to determine the best responses.
54
191740
2541
ื‘ื›ื“ื™ ืœืงื‘ื•ืข ืืช ื”ืชืฉื•ื‘ื•ืช ื”ื˜ื•ื‘ื•ืช ื‘ื™ื•ืชืจ.
03:14
Some also store memories of previous conversations
55
194281
3249
ื›ืžื” ื’ื ืื’ืจื• ื–ื›ืจื•ื ื•ืช ืฉืœ ืฉื™ื—ื•ืช ืงื•ื“ืžื•ืช
03:17
in order to improve over time.
56
197530
3385
ื›ื“ื™ ืฉื™ื•ื›ืœื• ืœื”ืฉืชืคืจ ื‘ืžืฉืš ื”ื–ืžืŸ.
03:20
But while Cleverbot's individual responses can sound incredibly human,
57
200915
4151
ืื‘ืœ ื‘ืขื•ื“ ืชื’ื•ื‘ื•ืช ื‘ื•ื“ื“ื•ืช ืฉืœ ืชื•ื›ื ืช ืงืœื‘ืจื‘ื•ื˜ ื™ื›ื•ืœื•ืช ืœื”ืฉืžืข ืื ื•ืฉื™ื•ืช ืœื”ืคืœื™ื,
03:25
its lack of a consistent personality
58
205066
2026
ื—ื•ืกืจ ื”ืขืงื‘ื™ื•ืช ืฉืœ ืื™ืฉื™ื•ืช ื”ืชื•ื›ื ื”
03:27
and inability to deal with brand new topics
59
207092
3153
ื•ื—ื•ืกืจ ื™ื›ื•ืœืชื” ืœื”ืชืžื•ื“ื“ ืขื ื ื•ืฉืื™ื ื—ื“ืฉื™ื
03:30
are a dead giveaway.
60
210245
2613
ื”ืกื’ื™ืจื• ืื•ืชื” ืžื™ื™ื“ื™ืช.
03:32
Who in Turing's day could have predicted that today's computers
61
212858
3254
ืžื™ ื‘ื™ืžื™ ื˜ื•ืจื™ื ื’ ื™ื›ื•ืœ ื”ื™ื” ืœืฆืคื•ืช ืฉื”ืžื—ืฉื‘ื™ื ืฉืœ ื”ื™ื•ื
03:36
would be able to pilot spacecraft,
62
216112
2089
ื™ื”ื™ื• ืžืกื•ื’ืœื™ื ืœื ื•ื•ื˜ ื—ืœืœื™ืช,
03:38
perform delicate surgeries,
63
218201
2317
ืœื‘ืฆืข ื ื™ืชื•ื—ื™ื ืขื“ื™ื ื™ื,
03:40
and solve massive equations,
64
220518
2289
ื•ืœืคืชื•ืจ ืžืฉื•ื•ืื•ืช ืžื•ืจื›ื‘ื•ืช,
03:42
but still struggle with the most basic small talk?
65
222807
3503
ืื‘ืœ ืขื“ื™ื™ืŸ ืœื”ืชืงืฉื•ืช ื‘ื ื™ื”ื•ืœ ืฉื™ื—ืช ื—ื•ืœื™ืŸ ืคืฉื•ื˜ื”?
03:46
Human language turns out to be an amazingly complex phenomenon
66
226310
3509
ืžืกืชื‘ืจ ืฉื”ืฉืคื” ื”ืื ื•ืฉื™ืช ื”ื™ื ืชื•ืคืขื” ืžื•ืจื›ื‘ืช ืœื”ืคืœื™ื
03:49
that can't be captured by even the largest dictionary.
67
229819
3705
ืฉืืคื™ืœื• ื”ืžื™ืœื•ืŸ ื”ื’ื“ื•ืœ ื‘ื™ื•ืชืจ ืœื ืžืกื•ื’ืœ ืœืฉื—ื–ืจ.
03:53
Chatbots can be baffled by simple pauses, like "umm..."
68
233524
4469
ืจื•ื‘ื•ื˜ื™ื ืžื“ื‘ืจื™ื ื™ื›ื•ืœื™ื ืœื”ืชื‘ืœื‘ืœ ืžื”ืคืกืงื•ืช ืคืฉื•ื˜ื•ืช, ื›ืžื• "ืื”ื”ื”...."
03:57
or questions with no correct answer.
69
237993
2422
ืื• ืžืฉืืœื•ืช ืฉืื™ืŸ ืขื‘ื•ืจืŸ ืชืฉื•ื‘ื” ื ื›ื•ื ื”.
04:00
And a simple conversational sentence,
70
240415
1900
ื•ืžืฉืคื˜ ืฉื™ื— ืคืฉื•ื˜,
04:02
like, "I took the juice out of the fridge and gave it to him,
71
242315
3280
ื›ืžื• "ื”ื•ืฆืืชื™ ืืช ื”ืžื™ืฅ ืžื”ืžืงืจืจ ื•ื ืชืชื™ ืœื• ืื•ืชื•,
04:05
but forgot to check the date,"
72
245595
1811
ืื‘ืœ ืฉื›ื—ืชื™ ืœื‘ื“ื•ืง ืืช ื”ืชืืจื™ืš,"
04:07
requires a wealth of underlying knowledge and intuition to parse.
73
247406
5466
ื“ื•ืจืฉ ื™ื“ืข ืžื•ืงื“ื ืขืฉื™ืจ ื•ืื™ื ื˜ื•ืื™ืฆื” ื›ื“ื™ ืœื”ื‘ื™ื ื•.
04:12
It turns out that simulating a human conversation
74
252872
2631
ืžืกืชื‘ืจ ืฉื”ื“ืžื™ื™ืช ืฉื™ื—ื” ืื ื•ืฉื™ืช
04:15
takes more than just increasing memory and processing power,
75
255503
3609
ื“ื•ืจืฉืช ื™ื•ืชืจ ืžื–ื™ื›ืจื•ืŸ ื•ื›ื•ื— ืขื™ื‘ื•ื“ ืžื•ื’ื‘ืจื™ื.
04:19
and as we get closer to Turing's goal,
76
259112
2511
ื•ื‘ืขื•ื“ ืื ื• ืžืชืงืจื‘ื™ื ืœืžื˜ืจื” ืฉืœ ื˜ื•ืจื™ื ื’,
04:21
we may have to deal with all those big questions about consciousness after all.
77
261623
4860
ืื•ืœื™ ื‘ื›ืœ ื–ืืช ื ืืœืฅ ืœื”ืชืžื•ื“ื“ ืขื ื›ืœ ื”ืฉืืœื•ืช ื”ื’ื“ื•ืœื•ืช ื‘ื“ื‘ืจ ืชื•ื“ืขื”.
ืขืœ ืืชืจ ื–ื”

ืืชืจ ื–ื” ื™ืฆื™ื’ ื‘ืคื ื™ื›ื ืกืจื˜ื•ื ื™ YouTube ื”ืžื•ืขื™ืœื™ื ืœืœื™ืžื•ื“ ืื ื’ืœื™ืช. ืชื•ื›ืœื• ืœืจืื•ืช ืฉื™ืขื•ืจื™ ืื ื’ืœื™ืช ื”ืžื•ืขื‘ืจื™ื ืขืœ ื™ื“ื™ ืžื•ืจื™ื ืžื”ืฉื•ืจื” ื”ืจืืฉื•ื ื” ืžืจื—ื‘ื™ ื”ืขื•ืœื. ืœื—ืฅ ืคืขืžื™ื™ื ืขืœ ื”ื›ืชื•ื‘ื™ื•ืช ื‘ืื ื’ืœื™ืช ื”ืžื•ืฆื’ื•ืช ื‘ื›ืœ ื“ืฃ ื•ื™ื“ืื• ื›ื“ื™ ืœื”ืคืขื™ืœ ืืช ื”ืกืจื˜ื•ืŸ ืžืฉื. ื”ื›ืชื•ื‘ื™ื•ืช ื’ื•ืœืœื•ืช ื‘ืกื ื›ืจื•ืŸ ืขื ื”ืคืขืœืช ื”ื•ื•ื™ื“ืื•. ืื ื™ืฉ ืœืš ื”ืขืจื•ืช ืื• ื‘ืงืฉื•ืช, ืื ื ืฆื•ืจ ืื™ืชื ื• ืงืฉืจ ื‘ืืžืฆืขื•ืช ื˜ื•ืคืก ื™ืฆื™ืจืช ืงืฉืจ ื–ื”.

https://forms.gle/WvT1wiN1qDtmnspy7