Will Superintelligence Lead to Spiritual Enhancement?

Round 1

Reviewer 1 Report

This paper is an excellent survey of a wide variety of authors working in and around AI and H+.  The author's thesis is clear, exceptionally widely supported, and well organized.  If anything, I'd say the paper verges on quoting too many sources, occasionally at the expense of the author's own insights.  However their voice comes through clearly.

I was distracted by the journalistic style of so many short paragraphs, some of them only one or two sentences long.  I'd suggest combining those that speak to the same general topic (such as 116-121, 127-135, 228-234, 288,300, 321-324, etc.).

The paragraph 223-227 seemed out of place as 228 flows from 222.  Omit?

The one section I'd have liked a bit more in was 5.3.  Why does relationship need a body?  A bit more said on this would help.

A few typos:  31 neighborly love, 151 highly developed, 153 the intelligence lane, 212 disembodied.

Otherwise, an important contribution to the discussion! 

Author Response

Number of short paragraphs reduced. I kept some of the short paragraphs which introduce a new topic.

Why do theologians need an embodied intelligence? I've added justification via a biblical passage, 1 Corinthians 15: 42-44.

Thank you for pointing out typos.

 

 

Reviewer 2 Report

Review of Spiritual Enhancement through Superintelligence? Really?
Rhetoric tone.  Beginning with the title itself, the rhetorical tone is dismissive.  The paragraph that begins on line 24 is probably the worst, but the tone throughout seems condescending in a way that is not appropriate for an academic journal. Often claims are made by assertion, not by argument.
Core Argument.  The abstract suggests that the core claim is that intelligence (a) lacks a moral compass and (b) does not include love, and therefore any enhancement of intelligence will not “count as spiritual advance” as Christians understand the term. 
Some Questions and Suggestions.
Line 20: Synthesizes?  No major theologians claims that God’s will and human will are synthesized.  Human will is conformed to, obedient to, and surrendered to God’s (eg Gethsemane prayer and the Lord’s Prayer). 
Perhaps here in the Intro, the AU could offer a definition of spirituality as it is being used in the article. Several terms are used--faith, love of God, love of neighbor, sanctification (please define), “Christian flourishing.”  Dimensions like prayer, meditation, contemplation, awe, disciplines and practices, are not mentioned.  Should they be?
In line 28 and following, reference is made to “human intelligence” and then to artificial intelligence.  Which are we talking about enhancing here? Enhanced human intelligence by means of AI and other technologies, such as genetics, so that our children will be more intelligent than we are (see line 249)?  Or smart machines?  Maybe both, but it’s not always clear.  It would seem that human intelligence and its possible enhancement must be front and center here because of the way the overall argument is framed: Will increased human intelligence count as increased spiritual intelligence?  The argument is not: will AI make the universe more spiritual.  (Kurzweil and some others could be seen as making this claim, but this article doesn’t seem to take that on.)  Its focus is on human intelligence and human spirituality. 
Line 60, following a long quote from Bostrom, basically accuses him of focusing on “one human trait and one trait only, namely, intelligence.” I’m not sure that’s fair to Bostrom.  It’s certainly not fair if we look at human enhancement literature more broadly.  To say there’s nothing here on love or “tenderness” is simply wrong. There is a rich literature on moral bioenhancement that is almost entirely ignored by the AU (more later on this). 
Section 3 (line 81ff) seems like nothing more than a long quote or recycled material.
Section 4 asks a really important question, central I would think to the coherence of the whole argument.  Unfortunately the brief section tells us more about ice cream than  about agape.  This section and the previous one need major revisions
Section 5.  In contrast to section four, which runs for 11 lines, Section 5 runs for more than 200.  Odd structuring.  Some specific points in Section 5:
The use of Homo sapiens sapiens is kind of dated and the timeline given doesn’t stand up well to the latest research.  
Then there follows a long and complicated discussion on human rationality and whether it’s unique to our species.  How does this relate to the core argument of the chapter? Lines 146-148 seems to imply that it does not relate.   
The list of seven traits (line 174ff) is helpful.  In 5.2, however, an interesting claim is made: AI lacks trait 7 and is therefore “not intelligent at all.”  Does this contradict the previous section in which it is claimed that various forms of intelligence exist, and most of them do not include trait 7?  Maybe the AU intends to say AI is not intelligent in the way humans are.  One point made here is that AI is easily fooled.  So the obvious question arises: When we humans are easily followed and fail to show “sound judgment,” have we lost our intelligence? 
But then 5.4 (273ff) goes into a long argument about technologies and their destructive powers.  The point about technologies as “neutral tools” or “morally benign” (!) seems a little naive.  But the main problem here is that human inventions have been used badly, which seems to call into question whether humans actually possess trait 7, or “sound judgment.” In fact, the article makes a pretty strong case against our ability to show sound judgment (so easily fooled, so gullible, deceptive and so willingly deceived…etc.). Then the AU worries that AI will make more powerful tools while lacking sound judgment.  To me, that sounds like AI will be a lot like intelligent humans, smart but dangerous.  So which is it: AI is not intelligent because it is not like us, or it is dangerous because it is a lot like us?   
Section 6 raises an interesting question.  One point might be clarified, however.  What exactly do we mean by the Ultimate? The Gilkey and Willows quotes imply that Ultimate=God. But then lines 336ff say that Love is the ultimate.  Maybe God=Love?  But that’s not how love is being used here.  It’s human love, with the claim that love is a higher end than intelligence.  
Line 368ff raises a key question: Could any technology—whether ML, AI, IA, or genetic engineering enhance the human capacity for loving? For virtuous living? For sanctification or deification?”  In some ways, this seems to be the central question of the essay, ignored unfortunately up to this point.  If the core argument is that increased intelligence does not equal increased spirituality, that claim is pretty made.  If however the core argument is that technology cannot increase any dimension or feature of human spirituality, that raises a far more difficult challenge.  The AU seems to ignore the question of moral bioenhancement, at least up to this point. Is this article then a straw person exercise? Tech might increase intelligence but that won’t increase spirituality. That was easy, but does anyone really claim otherwise?  The real challenge is that tech might increase components of spirituality, depending of course on how that’s defined.
Section 6.1 is key, but unfortunately the analysis tends to limit itself to genetics, which is fairly easy to dismiss as a pathway to spiritual enhancement.  Other technologies are not so easily dismissed.  See Salvelescu and Persson, Harris Wiseman, or Brian Earp among others. Techs range from pharmaceuticals to brain stimulation to psychedelics.  Can these make humans more spiritual? It’s debatable, of course, but the AU seems to ignore the question except as applied to genetics.  

 

Author Response

Thank you to this peer reviewer for a review that is thorough and specific. Many of the editorial suggestions will improve the manuscript. Here I respond.

The rhetorical tone? I certainly wish to demonstrate respect to professionals in AI and transhumanism, so please notice the title change and rewriting of the mentioned passages. Specifically, line 24 is rewritten.

"Synthesizes" wills has been changed to conformity. Good theological point.

"Spirituality" definition added to introduction.

Ambiguity regarding "intelligence" reduced.

Yes, Nick Bostrom is more nuanced. Text now changed to acknowledge Bostrom's concerns for freedom and dignity.

Section 4 on Agape love now expanded beyond the ice cream analogy.

I believe the inclusion of the seven stages of intelligence is important for the subsequent arguments. This is because all intelligence we know to date is embodied. If AGI seeks to mimic human intelligence, then embodiment should be presupposed. This requires a burden of proof on the part of those pursuing AI Superintelligence to conceive of a post-biological host for intelligence. 

I added material to distinguish between judgment and sound judgment.

I'm sorry that you deem the moral ambiguity of technology to be naive. What I find problematic in transhumanism is the naivete that enhanced intelligence is an unambiguous good to be sought. Part of my point is that this naivete needs more transparency and the theologian can provide that transparency.

Yes, moral bioenhancement discussions are widespread. I've now contexualized my discussion of genetic enhancement within this wider discussion.

Overall, these criticisms have improved the article. Thanks.

 

 

Reviewer 3 Report

Wonderful article! Great insights concerning the nature of intelligence. Clearly presented, extremely engaging, and well-organized.

(Entirely Optional) Suggestions:

The following sentence, "We find it occasionally in the animal kingdom as well" (p. 3.105-06) might be better placed immediately below the block quote, as it its current location suggests that the block quote immediately following will be about the animal kingdom.

The sentence describing neighbor-love as inherent in humanity (p. 8.360-61) might be challenged in light of the reality of (universal) human sinfulness.

In the concluding sentence the author nicely states that "the love of the heart takes precedence over the genius of the mind." In order however to avoid perhaps an overly-sentimental view of the heart, it might be worth pointing out in a footnote that Christians (and Jews) are exhorted to love God with one's heart, mind and soul (Deut. 6; Matt. 22), however these might be understood. In other words, love (agape) is not merely an issue of the heart.

Typographical issues:

• footnote vii is not in superscript
• Page 4.151 "Our highly develop[ed] human reasoning . . . "
• Page 7.309 "e4ven wit immense effort"
• agape is not consistently italicized
• Page 13, end note 2: "brought about [by?] increasing remembrance of God"

Author Response

Reworded the sentence containing "animal kingdom," almost but not quite as you suggested.

On neighbor-love, added a dependent clause to acknowledge sin and the moral ideal.

Yes, Deut. 6 & Matt.22 include the mind. I put that in.

Thanks for the helpful critique.

 

 

Reviewer 4 Report

The manuscript is mainly a useful argument only within a Christian community whose discussion is uncritically based upon their own theological claims.  Different issues such as AI and Genetically Engineering are discussed together, that is only possible from their daily theological conversation.

Though a topic of what intelligence is important generally, authors do not know the history of scholarly arguments on what intelligence means in the AI field.   Nobody cares if Christian community is based upon love or not in the argument of AI.   There are much better theological studies on technology. This manuscript is not one of them.

The manuscript is not suitable for Religions. 

Author Response

Thanks for pointing out typos.

You think that "nobody cares if the Christian community is based on love or not in the argument of AI." Well, it appears to me that with the strong history of revering human reason within Christian theology, pointing to the preeminent place of love could have some contemporary value.

 

Round 2

Reviewer 2 Report

Thanks for taking the review seriously.  In my view, the result of the review process is very positive.