Monday, October 26, 2020

Erwin Questionnaire

 A guy I knew way back sent me a research questionnaire, so I obliged and filled it out. I don't feel especially qualified to respond but I figured what the heck. The questionnaire is in plain text, my responses in italics.

-----------------

My name is Rich Erwin and I’m a first year graduate student in the Foresight program at the University of Houston. I am working on a project regarding what might constitute a more human-centered Data Revolution, one which accounts more for individual needs of economic value, self-worth, and self-representation, which is defined within the questionnaire below.

My goal is to determine whether any trends exist among those who are among the leading creators, pathfinders, observers and analysts of this revolution in progress, to determine what ideas might need more visibility and consideration, and to potentially spark further discussion regarding the topic. 

A summary statement and nine  questions are provided below.  I would be very grateful for a moment of your time and consideration to answer each to the best of your ability and return the answers to me by 5:01 PM Pacific Time on Thursday, October 21st (12:01 AM UCT on Friday, October 22nd).

Any responses will remain within the context of my coursework.

If you have any questions or concerns, please do not hesitate in contacting me by email, text or voice.

Best Regards,

[REDACTED] 

 

Questionnaire

 

Humanity has had agricultural and manufacturing revolutions.  I believe that, with the maturing of the internet and access to relatively cheap and vast computational power, how both can be used together have placed us in the midst of a Data Revolution.  However, as in the previous two revolutions, we are dealing with significant issues of how we as individuals can and are allowed to react to it in terms of economic value, self-worth, and self-representation. 

 

I believe that you have a unique insight into this issue and would greatly appreciate your responses to the questions below. 

 

1)      What tools, processes or precepts do you believe need to be in place in order to drive, effect and maintain a more human-centered Data Revolution, and why?

 

 

A core element of being fully human is Autonomy: independence from external control. Becoming adult is a pleasure because it increases autonomy and jail is a punishment because decreases it.

Many or most Data Revolution (DR) technologies make gestures toward autonomy in self-representation, (e.g. select your Profile image, checkbox news feed interests); nourish your sense of self-worth (post content that earns “likes”, block people who annoy you); and impact your economic value to DR actors (respond to or block ads). These gestures seem doomed to failure in an environment with Big Data analysis of shopping patterns and cellphone data, stoplight cameras, and facial recognition applied to photos outside your control [Here’s an example: an American was vacationing in Cuba in violation of the embargo, and careful not to post vacation photos on the internet. An unrelated person posted vacation photos with that person in the background. The relevant American office used software to identify the American and fine them. One can imagine many variations on this theme.]

A truly human-centered DR would preserve human autonomy, and I have not the faintest idea how that might be accomplished. Perhaps --- like freedom itself --- it can only be a guiding principle to be striven for and improved upon, but never completely met.

 

 

 

 

 

2)      What advice or recommendations would you provide to a young adult regarding how to navigate their life in the midst of the Data Revolution?

 

 

Identify and rigidly control the Personally Identifiable Information (PII) that can hurt you if disclosed, and use cammo without apology. For example, many apps ask for your date of birth (DOB) that don’t really need it; often they are simply verifying your age. If they get hacked – or rather WHEN they get hacked – criminals will have a clue that’s helpful for robbing you, e.g. electronically filing your tax return and collecting your refund to a burner account. You need to camouflage your PII where possible. Pick a date – such as January 1 – that is the DOB you’ll give out to those apps asking to match your food preferences with your Star Trek character, and reserve your actual DOB for those very few institutions that need it, e.g. your federal tax accounts.
You can take similar precautions with other information: systematically misspell your name, use a po box not a street address, etc. This is all chaff, and can be defeated by a determined AI, but you can be less vulnerable.

 

There’s a tradeoff to all this of course. All that data about yourself out there can be used to build deeper relationships with people physically distant from you. Before DR, it was rare for friendships to survive a cross-country move – which is why 25 Year Class Reunions used to be a big deal. With DR, you can have a class reunion every Sunday. Embrace this.

 

 

 

3) How would you rate the issues below as impediments to a human-centered Data Revolution from most intractable (1) to least intractable (6)? (If you think any are not at all intractable, please mark with an X and explain why.)

 

            a)  Surveillance through access of “infinite content”

(5) This is an inherently intractable problem, since in a competitive world there is enormous benefit to an organization – whether geographic state or stateless actor – that surveils its members and non-members. For example, if we established an ideal scheme of controlling our government’s use of data, we have done nothing about actions by other governments and they would have a competitive edge against our government. The solution cannot be to give up but I don’t see the solution yet.

 

            b)  The effect of data algorithms in use on diversity, equity and empathy

(4) As a person with mostly majoritarian characteristics (white, male, grumpy) I am reluctant to comment on the impact of algorithms on diversity – I just don’t have the life experience to make evaluations – but I can speculate that it might be a mixed bag. Facial recognition is notoriously bad at identifying nonwhite faces, to the significant disadvantage of POCs in this month’s administration of bar exams across the nation ( https://abovethelaw.com/2020/09/online-bar-exams-rely-on-facial-recognition-tech-and-guess-what-its-still-racist/ )

Algorithms could be used to promote inequality (e.g. by identifying race based on matching job applicants to their Facebook photos) or equality (e.g. by sifting out actual job qualifications from irrelevancia) but I just don’t have enough experience to comment further.

I would like to see DR to do something about empathy. Online personas famously divorce a person’s humanity from their digital representations; the use of handles and nonhuman avatars only make it easier to forget that behind some of these online postings are actual people. I don’t see an easy solution.

 

            c)  Social Media and the byproducts of “truth versus popularity”

(6) Algoritmic-based selection of content seems to have the pernicious effect of presenting on the basis of popularity rather than quality (and never mind truth ….) because the most popular sites make money based on clicks. This is how they live: by offering content you’ll click on.

I think it’s wonderful – and surprising – that there are small efforts to doing something about that – such as twitter/facebook amending posts about voting with links to information about voting – but that doesn’t seem like enough. The care that facebook/twitter took with Giuliani’s ridiculous laptop story is praiseworthy but the pushback was enormous. I expect future scams to be a little more subtle because the payoff is immense.

 

 

            d)  Potential for conflict within society between individuals significantly reliant on virtual versus             physical reality

(3) As of yet I don’t see this conflict, but who knows?

 

            e)  Physical self-augmentation versus replacement by artificial intelligence

(2) I feel it most likely that AI and its junior cousin VR control of drones will swiftly outpace physical self-augmentation, because the latter has a weak point: the human body. For most tasks involving interaction with physical reality, Ais will be better at everything except taking sick days.

 

            f)  The effect of the Data Revolution on the eventual advent of the Genetic Revolution

            (CRISPR, TALEN, Epigenetic Editing, etc…), specifically with regard to restoration or prevention             versus the desire for enhancement

(X) While DR/GR interactions have grave potential for harm, they have plenty of opportunity for benefit as well so I don’t see them as intractable problems; ultimately there is a political question of who shall set up the rules for whose benefit

 

4) Which would you consider the most problematic issue regarding “infinite content”?

            a)  Surveillance by authorities or factions

***This has always been problematic, but “in theory” could be controlled by an alert citizenry with effective representatives. The theory does not always seem to have been borne out in practice but I retain hope.

            b)  Treatment of your data and actions online as monetizable content

***This is really annoying – I want a piece of the action – but not a direct harm to me, so I would rate it as least problematic. It reminds me of the biotech industry patenting genes taking from patients without compensation, e.g. the Henrietta Lacks scandal. I despair of any justice there.

            c)  Potential for attempted behavior modification
***This is the most problematic issue, since it directly impacts personal autonomy.

            d) Self-restriction of access to a diverse set of viewpoints over time
***This is problematic indeed, but it’s a self-inflicted wound and so not quite as bothersome as 4c. Still, we see the negative impacts of this self-restriction playing out in the current American elections as each party becomes more convinced the other is the Devil.

            e)  Potential for creation of artificial constructs of yourself
***I don’t see this as problematic to the extent that contructs would be controlled by myself. It could be helpful to agent myself for routine tasks, although collating the various experiences of my selfstream sounds like an interesting problem.  Constructs controlled by someone else sound creepy; I imagine state entities might want them to predict and modify my behavior (see 4c); I would like a ban on private entities doing the same.

            f)  All five are equally problematic
***Disagree

            g)  No such issue exists

***Disagree

 

5) By what standardized means should data algorithms be audited for potential issues regarding diversity, equity and empathy?

            a) Emphasis on frequency – at least once a year
*** Audits should be automated so they can be continuous, just as virus checking is continuous.

            b) Emphasis on auditing outside the entity that uses the algorithms in question
*** Internal audits are nice for an organization determined to avoid problems, but outside audits are necessary  for the same reason that we have virus checkers separate from the app provider.

            c) Emphasis on standardized governmental regulations that cover a) and b)
*** Governmental regulations set minimums that are (in theory) determined by the public (so there’s more confidence in them) but perhaps it is more important that they set up a level playing field, without which there can be a “race to the bottom” – incentives for cheating.

            d) Emphasis on insurance companies requiring audits in return for coverage against lawsuits
***In an analogous situation: some of the best Continuing Legal Education teaching lawyers to avoid ethical problems like to result in lawsuits comes from insurers (such as Attorney Protective) that provide them in an obvious and laudable attempt to minimize problems and payouts. IMO this works because any law to immunize lawyers from being sued would be a tough sell to an elected legislature hahaha!
OTOH DR firms have enough money to purchase a bill immunizing them; also they can word the standard user agreement to send things to arbitration (and arbitrators usually favor them), and if nothing else just stretch out litigation until it’s pointless (as the saying goes: “Delay, Deny, Until They Die”). TL;DR: lawsuits may be nice but are unlikely to be an answer.

 

6) Given the state of social media, would you allow someone you care about to be introduced to it?

            a)  No

            b)  Yes, but only under very strict regulations, controls and rules

            c)  Yes, but only after educating them to improve their awareness, and with some controls and

                        rules in place

            d)  Yes, but only after educating them to improve their awareness

            e)  Yes


It depends on my relationship to them; if they’re some one I’m responsible for (e.g. child) then “C”; if someone for whom I have no responsibility other than friendship then “D”.

7) Would you support increased surveillance that focused on significantly reducing cyber-bullying?

            a)  No

            b)  Yes, but under very strict laws, controls and rules

            c)  Yes, but accompanied by extensive public education of the process

            d)  Yes

***I do not know enough about this subject to form an opinion. My uninformed opinion is that DR platforms should enforce Do-Not-Contact requests, so that if A is bullying B, then B can block A without any need to give a reason. I don’t know whether that would solve the problem.

 

8) Do you think people who spend significant time in virtual realities might eventually become a potential threat to physical reality?

            a)  No

            b)  No – people will always perceive what they want to perceive as their dominant reality

            c)  Yes – some people aren’t able to put each reality in perspective consistently

            d)  Yes – and we need laws to counter the threat

ALL human perception of reality is a virtual reality (VR). Our field of vision is limited and we are constantly filling in the gaps depending on what we looked at last. Many car accidents result from this sort of thing; that’s why many side mirrors have a printed warning “Objects In Mirror Are Closer Than They Appear”.

What I believe this question is referring to is electronically supported VRs which take things to new levels of divorcing our human perceptions and reality creation from what exists in the physical world.  People staring at their cell phones have the perception that they are aware of the road because their brains are filling in the gaps, right up to the crash. Drone operators thinking they are attacking enemy troops end up killing civilians. As long as VRs are connected to real-world tools, the opportunity for threats exist. (See John Scalzi’s “Lock-In” series for a light – but not light-hearted – examination.)
If the question concerns VRs not connected to physical tools, perhaps the danger is less. There is still the risk of harm from hacking data – but that’s not unique to VR – and self-harm from neglect of physical reality – but again that’s not unique to VR, IMO.

 

9) If your employment required physical augmentation of your body versus being replaced by an artificial intelligence construct, would you do so?

            a) Yes

            b) Yes, if who I worked for covered the cost of installation

            c) Yes, if who I worked for covered the cost of instillation, replacement and eventual removal

            d) Yes, if extensive regulations for my protection were in place

            e) No

 

If I was going to be paid for the work of my AI self, I’m cool with that. I assume that the question is instead that the employer gives the choice to take the augmentation or lose my job. No reasonable person would not trust an employer’s representation as to the safety of the augment; a third part would have to certify its safety.

(In this context the “SeaWorld v Perez” case is instructive: SeaWorld told its employee Dawn Brancheau that its working conditions were safe but they in fact were not; while the majority upheld OSHA’s authority to require SeaWorld to do something about that, the dissenting Judge Kavanaugh argued that the employee had assumed the risk. Kavanaugh is not on the Supreme Court. https://caselaw.findlaw.com/us-dc-circuit/1663286.html)

The nature of the augmentation and its reversability would be elements of my choice.


No comments: