#24 - Biologist: Some computer scientists imagine that all

[15veries]

Hi

So I see how B is correct but not sure how to eliminate A.
So the assumption I thought would be something about the protein and its interaction.
In the stimulus suggest the scientists think if they can manage info in genome, they can make artificial intelligence but the author says this is not the case, so it suggests the interaction of protein is more influential than genome itself...so even though they are encoded in the human genome, scientists would not be able to encapsulates info from studying genome. So thus B right?
But I felt A is very tempting too...it does not talk about protein's interactiton or genome directly neither, but it does kind of generalize the idea talked in B,
Is A wrong because it's too broad?

[lorettan102]

15Veries:

I did the same thing you did. In looking back at the stimulus and answers now I think B better explains why creating a computer program that encapsulates the the information in the human genome would not be suffice in making an artificial intelligence. Answer choices A and B seem very similar to me and I did choose A. However, as I said I think B pretty much nails it on the head, especially with the specific language shared in the answer and the stimulus.

Either way, I am curious to see what someone else has to say about this question.

[Jonathan Evans]

Hi, Lorettan102 and 15veries,

Great questions. This is a difficult problem and requires a lot of precision about what is absolutely necessary for the conclusion to be valid.

The claim is that: There is more required to making an artificial intelligence than creating a computer program that includes all information in the human genome.

The support offered is that: The operations of the brain are governed by the interactions of proteins, of which the structures are encoded in the genome.

For the claim to be valid, it must be true that these interactions cannot be replicated only through a computer program that includes all the genome information. This is Answer Choice (B).

Answer Choice (A) would certainly strengthen the argument, but it is not necessary for the argument to be valid. Consider the negation test. What if the functions of the human brain were governed by processes that could be simulated by a computer? It could still be the case that only including the information in the genome would be insufficient to determine the interaction of these proteins. Thus, the conclusion could still be valid, and Answer Choice (A) does not describe a necessary assumption.

[lorettan102]

Thank you!

[MBG13]

I would like to provide my thinking on answering this question. I must admit this the first assumption question that answered using a prephase or thinking like this, so it was pretty exciting.

P1 - Some computer scientist think all you have to do to create an artificial intelligence is "encapsulate the information contained in the human genome." (Here I thought, okay...this theory is about to get debunked.)

C - They are mistake! (Aha! I was right (emphasis is all mine)).

So this is where I began to have a real conversation with myself and asked: why are they mistaken? That lead to premise 2.

P2 - because a brain's "operation" is "governed by the interactions of proteins whose structures are encoded in the human genome."

So I thought, okay the differences between P1 and P2 is interactions. I almost went down the rabbit hole because I began to think.... well if the information is encapsulated into the AI...then wouldn't that include information about the proteins too.... That's when it hit me! Wait...they said they can encapsulate the information -- not that they can make it work (interact) the same way!

That's when I came up with the following loose prephrase: AI cannot encapsulate interactions of proteins! And Bam! That lead me to (B).

Thank you PS...two days before my test and I happy that I was able to do this without too much of a struggle!

By the way...when I was reading answers A and C, they struck me as almost saying the same thing (Function vs. Operation). Am I correct? I only noticed it on my second pass after I answered C.

[Emily Haney-Caron]

Hi MBG13,

I'm so glad to see that everything is really solidifying for you right before the test!

To answer your question, A and C are somewhat different. A is almost saying you can't quite create AI (human brain can't be simulated); C is saying you can make AI but you'd have to model it on the human brain.

[MBG13]

Okay...thanks for the response!!!

[nky45gho]

Hi, I'm having trouble understanding why C is incorrect. When I look at it, I attempted the assumption negation technique and came up with the following thinking:

If the only way to create AI is through modeling the operation of the human brain, then you'd have to model the interaction of those proteins. Since just their structures are encoded (and not their operations), then the computer scientists' model mentioned does not model the human brain and therefore fails. This seems to support the argument.

If just one way (but not the only way) to create AI is modeling the operation of the human brain, then it is possible the human genome is sufficient (and not the interaction of proteins). This seems to negate the argument. If there are other ways outside of the human brain, then it doesn't really matter that the interaction of proteins isn't represented.

Am I reading too much into how interaction of proteins and them being encoded equates to modeling the brain? I see how B passes the assumption negation test but I don't understand why it's stronger than C. Thanks for any help!

[Luke Haqq]

Hi nky45gho!

I'd be happy to address why (C) is incorrect, using the assumption negation technique.

That answer choice states,

(C) The only way to create an artificial intelligence is to model it on the operation of the human brain.

This is saying there's only one way. The negation states that there's not just only one:

(C) [It is not the case that] the only way to create an artificial intelligence is to model it on the operation of the human brain.

The second step of the Assumption Negation technique is to put this back into the argument and see if it falls apart. Effectively the negation is saying that there are other ways. But even if this were true, the biologist's conclusion could still be right that "They are mistaken."

Indeed, it's worth emphasizing one of the words in this very short conclusion--"they." This referring to the "some computer scientists" with certain beliefs in the prior sentence. The argument is about what they believe about AI development, rather than all actual possibilities about AI development. Those computer scientists believe all that is required for AI is a program that can encapsulate information of the human genome, but they are wrong because this alone would be insufficient. They're still wrong in that specific belief about AI development even if there are actually other ways to develop it. Thus the negation of answer choice (C) doesn't make the argument fall apart.

[lemonade42]

So are we supposed to assume that the biologist wants to create an AI that mimics the human brain? At first, I was thinking why would the author suddenly mention the brain when he was just talking about AI. I just thought of AI as a tool that would incorporate data from the human genome, not necessarily also have the need to be like the brain (I didn't want to bring in any "outside info" in assuming AI mimics the brain, so I just thought of it as a piece of technology). But since the answer is related to how we would need the interactions in the brain in addition to the human genome info, then wouldn't that mean we have to think that the biology wants to create an AI that mimics the human brain?