2
$\begingroup$

A few forces seem to dominate the determination of AI and cybernetic research direction.

  1. The quest to automate repetitive tasks
  2. Interest in the nature of consciousness, learning, and adapting
  3. Interest in language automation (librarian, translator, therapist, loyal friend)
  4. Outsmarting the foes in geopolitics
  5. Outsmarting the competitors in local and global economics

Which of these five motivations produce an impact on the direction of technology efforts that lacking key elements in risk management?

These are the corresponding five results one might expect upon success in each of the above five objectives.

  1. More human leisure time to waste on shopping, gossiping, entertainment, blogging without defined purpose, pretending to do vital office work eight hours a day when producing only an hour or two of usable work product.
  2. Clearer understanding of ourselves and our place and duty as a species.
  3. Cyborexia (fear of talking with our own species unless a computer creates a layer of safety in between, a phenomena already occurring) AND the entrusting to computers not only the marionette strings of friendship and trust but also the key relational bridge between cultures
  4. The development of tools intended to outsmart foes outsmarting all humans instead
  5. The development of tools intended to outsmart competitors outsmarting all humans instead

Only one of these five seems to be a humanity enhancer with low risk. The others seem to present a high risk to human destiny and offer no particular advancement of the human species.

AI is cast by futurists as the edge of advancement and demonized by technophobes. Can the development of AI be pointed more toward those objectives that will improve the world?

Does anyone see a second one or have a sixth or seventh force and corresponding result to offer? Is there any flaw in what appears to me to be obvious eventualities?

$\endgroup$
2
$\begingroup$

I tend to take a reductionist view and see economic incentives as the overwhelming driver of AI research, without regard to consequences.

I see the entire field as based on optimization, which includes automation of repetitive tasks. It's about reducing cost/downside and increasing return-on-investment in any area to which intelligence can be applied, such as the 5 examples you provide.

Even the idea that AI research can be oriented toward "interest in the nature of consciousness, learning, and adapting" allows us to manage humans and artificial systems more efficiently, and is ultimately utilitarian.

The philosophy of neo-luddism becomes increasingly important because of the "law of unintended consequences", which has been strongly validated in regard to technological advancement, certainly since the industrial revolution. (Mythologically it might be said to go back to the theft of fire by Prometheus--the technology carries great benefits but also destructive capacity.)

It seems pretty clear that we can only make assumptions about where it's leading and what the effects may be, whether good or ill, but that there will be perils is a certainty because we're already starting to see some of the negative outcomes in regard to algorithms optimizing social media engagement & bots used in nefarious ways.

$\endgroup$
0
$\begingroup$

A few forces seem to dominate the determination of AI and cybernetic research direction.

  1. The quest to automate repetitive tasks
  2. Interest in the nature of consciousness, learning, and adapting
  3. Interest in language automation (librarian, translator, therapist, loyal friend)
  4. Outsmarting the foes in geopolitics
  5. Outsmarting the competitors in local and global economics

Where does this list come from? It seems rather arbitrary to me;

  • Point 3 could be viewed as a part of point 1.
  • Points 4 and 5 have overly restrictive (and probably overly negative) phrasing, which means that 1) the list is certainly not complete, and 2) the phrasing turns the entire question into kind of a leading question.

I'd say the list would be a much better reflection of reality if points 4 and 5 were collapsed into a single point:

  1. Improving (decision-making) processes

In some cases this can indeed be about outsmarting competitors in various senses of the word (e.g. doing something more efficiently than competitors, allowing you to sell products/services at a cheaper rate). There are plenty of non-competitive settings as well though (e.g. optimizing processes in healthcare). Again, this point can have overlap with the first point (of automation), though it doesn't always have to be about automating things that are currently done manually; it can also be about doing things that we weren't able to do manually in the first place.

First Question:

Does anyone else see four out of five of these motivations as somewhat lacking in intelligence?

This is difficult to answer precisely without first attempting to dive into a precise definition of "intelligence". In broad terms though, I think this is an artifact of the overly negative original phrasing of the "forces". I personally think it can certainly be intelligent to strive for better automation, or improved (decision-making) processes, or better translations, and all of those tasks may also require varying degrees of intelligence.


Based on these goals, these are the corresponding five results one might expect upon success.

  1. More human leisure time to waste in pointless shopping, gossiping, entertainment, pretending to do vital office work eight hours a day, and blogging than even today.
  2. Clearer understanding of ourselves and our place and duty as a species.
  3. Cyborexia (fear of talking with our own species unless a computer creates a layer of safety in between, a phenomena already occurring) AND the entrusting to computers not only the marionette strings of friendship and trust but also the key relational bridge between cultures
  4. The tools of outsmarting outsmarting us (casting us as geopolitical foes)
  5. The tools of outsmarting outsmarting us (casting us as competitors)

Second Question:

I only see one intelligent objective among the five. Does anyone see a second one or have a sixth or seventh force and corresponding result to offer? Is there any flaw in what appears to me to be obvious eventualities?

  1. That again looks like overly negative phrasing to me. It's likely that there will be plenty of domains where we don't want (law enforcement, healthcare, etc.) or simply don't manage (e.g., teaching) to achieve 100% automation. In many cases, we'll still have jobs to cooperate with automation, where we need to consistently and carefully watch to make sure nothing breaks down, perform maintenance, etc. And even if somehow all of that can truly be automated 100%, many people would argue that the gained free time will not be "wasted" or "pointless".

  2. No comments, I guess this is the one you already judged to be "intelligent".

  3. Doesn't seem like an obvious eventuality to me at all. Can't find any literature on it either by searching for that term at least since it looks like you came up with it and are the only one using it? I don't see why "fear of talking with others" would be a natural consequence of advances in chatbots, automated translation, or any other form of NLP. It just opens up more opportunities to communicate with different people and/or in different ways. Related food for thought: did the invention of the telephone make people afraid of talking in-person?

Points 4 and 5 would naturally, in my opinion, be replaced by:

  1. The tools for improved (more efficient in terms of energy/time/money, more performant in terms of results/outcomes) (decision-making) processes

which certainly seems like a worthwhile objective to me (whether or not that also means it's an intelligent objective depends on definitions).

$\endgroup$
  • $\begingroup$ @DouglasDaseeco Along those lines; yes, AI research has the potential to lead to various "bad" results. I don't think that's unique to AI research though, that has been a common theme in science throughout the centuries. If humans never figured out how to create tools from iron, we'd also have fewer weapons. If there were fewer advances in biology/chemistry/physics, we wouldn't have poisons/chemical weapons/nukes. It's certainly wise to keep potential negative effects in mind, try to avoid them. There are also many positive effects though, and avoiding negatives won't always be possible. $\endgroup$ – Dennis Soemers Jul 29 '18 at 16:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.