> “The Turing test started as a way to ask whether machines could rival human intelligence”
Turing never proposed this in his paper...
He wrote:
//If the meaning of the words "machine" and "think" are to be found by examining how they are commonly used it is difficult to escape the conclusion that the meaning and the answer to the question, "Can machines think?" is to be sought in a
statistical survey such as a Gallup poll. But this is absurd.//
Implicit to his observation is the stipulation that the only reference for thought available is what humans do, so he proposed a placeholder test called the imitation game, and proposed criteria for interrogation, including a gender-pertinent test configuration.
And just as salient to the the test is what constitutes a "machine".
Well, this report totally avoids consideration of Turing's test criteria even as it uses his name.
By reducing "thinking" and "machine" to the ordinary meanings of these words and observing humans who interact in thoughtful discourse with the device in a natural way, it looks to me like a generic Turing test was passed by Weisebaum's Eliza (The Doctor) in the 1970s. The conclusion drawn that it's not that difficult to seem human, and we find examples of anthropomorphization all around us.
If you read his paper with attention to tone, Turing might be flabbergasted at the simulation capabilities of today's AI, but he anticipated such a development.
But this report doesn't really consider Turing's test as he proposed it, and seems wholly uninterested in Turings questions about what it means to think and what constitutes a machine.
> “The Turing test started as a way to ask whether machines could rival human intelligence”
Turing never proposed this in his paper...
He wrote:
//If the meaning of the words "machine" and "think" are to be found by examining how they are commonly used it is difficult to escape the conclusion that the meaning and the answer to the question, "Can machines think?" is to be sought in a statistical survey such as a Gallup poll. But this is absurd.//
Implicit to his observation is the stipulation that the only reference for thought available is what humans do, so he proposed a placeholder test called the imitation game, and proposed criteria for interrogation, including a gender-pertinent test configuration.
And just as salient to the the test is what constitutes a "machine".
Well, this report totally avoids consideration of Turing's test criteria even as it uses his name.
By reducing "thinking" and "machine" to the ordinary meanings of these words and observing humans who interact in thoughtful discourse with the device in a natural way, it looks to me like a generic Turing test was passed by Weisebaum's Eliza (The Doctor) in the 1970s. The conclusion drawn that it's not that difficult to seem human, and we find examples of anthropomorphization all around us.
If you read his paper with attention to tone, Turing might be flabbergasted at the simulation capabilities of today's AI, but he anticipated such a development.
But this report doesn't really consider Turing's test as he proposed it, and seems wholly uninterested in Turings questions about what it means to think and what constitutes a machine.
> [...] with the right “persona” prompt, advanced AI can pass as human in live chats.
> Without persona prompting, performance dropped sharply.
Suggestion: Compare the prompted AI's to human con artists, lobbyists, and sycophants.
[flagged]