Monday, April 24, 2017

Systemic / Structural Bias / Prejudice and Privilege Embedded in Software and now Artificial Intelligence

Good morning.
"And the beat goes on...................."

When we talk about the concepts of racism and privilege being structural and systemic, we mean that the biases and prejudices, and perks and advantages are embedded in the structures and systems by which we operate and the environments in which we live.  Those systemic realities transcend individuals, situations and time, and are often invisible in the way they insinuate into our cultures and influence who we are, our attitudes and beliefs, and our decision making processes.

Everyone is to some degree prisoner to their heritages, histories, cultures and environments, and biased and prejudiced accordingly.  Many, if not most, people are largely unaware of the extent to which they have been programmed, and how those programs perpetuate social and political climates and how we function daily.  All best intentions aside, that makes it axiomatically more difficult to address the inequity challenges facing the world.

I have been interested of late in the dangers in the growth of Artificial Intelligence (AI) and how machine learning ultimately threatens the existence of the human species as machines elevate to the superior position in the relationship with humans - greater information and knowledge, and its own learned kind of wisdom based on precepts and assumptions it (not we) may make; leading to the day when the machines tire of the human species and see no logical reason for its continued existence.  Science fiction? Maybe, but there are lots of minds out there who share that fear.   Moreover,  in the short term machines - being programmed by human beings - are likely to exhibit the same biases and prejudices as those who programmed them - thus perpetuating the inequities of modern life and the myriad problems stemming therefrom.

A recent article in the Guardian reported on troubling research that AI Programs exhibit racial and gender biases and prejudices:


"The findings raise the spectre of existing social inequalities and prejudices being reinforced in new and unpredictable ways as an increasing number of decisions affecting our everyday lives are ceded to automatons.
As machines are getting closer to acquiring human-like language abilities, they are also absorbing the deeply ingrained biases concealed within the patterns of language use, the latest research reveals.
Joanna Bryson, a computer scientist at the University of Bath and a co-author, said: “A lot of people are saying this is showing that AI is prejudiced. No. This is showing we’re prejudiced and that AI is learning it.”
Bryson warned that AI has the potential to reinforce existing biases because, unlike humans, algorithms may be unequipped to consciously counteract learned biases."


Developments in AI are happening at an accelerated rate.  This isn't science fiction way in the future stuff - this is happening now, and likely to happen ever faster.

This made me wonder to what extent (and to what damage) software that has guided our computational efforts for the past three or more decades has already been embedded with bias and prejudice?  How has privilege already been incorporated into all of the "smart" devices that now manage and empower our lives?  How many online video games that kids play millions of times a month, unintentionally reflect the human biases and prejudices and privileged positions of those who created the games - manifested in language, preferences, rewards and otherwise?  And what is the net effect of that?  How many software programs created to power the search engines we use, reflect that same unintentional bias?  How many myriad ways we use computer programs and software have further increased and deepened the structural and systemic racial, gender, sexual orientation, religious and other prejudices with which we still grabble.

The paper reported on in the Guardian above pointed out:


"The latest paper shows that some more troubling implicit biases seen in human psychology experiments are also readily acquired by algorithms. The words “female” and “woman” were more closely associated with arts and humanities occupations and with the home, while “male” and “man” were closer to maths and engineering professions.
And the AI system was more likely to associate European American names with pleasant words such as “gift” or “happy”, while African American names were more commonly associated with unpleasant words.
The findings suggest that algorithms have acquired the same biases that lead people (in the UK and US, at least) to match pleasant words and white faces in implicit association tests.
These biases can have a profound impact on human behaviour. One previous study showed that an identical CV is 50% more likely to result in an interview invitation if the candidate’s name is European American than if it is African American. The latest results suggest that algorithms, unless explicitly programmed to address this, will be riddled with the same social prejudices."


So if the problem isn't just the algorithms that will facilitate the learning by AI, but are already working their negative impacts in software and programs we use and have been using, we (in our own field of the nonprofit arts) need to try to figure out which programs, which software specifically - and in which situations - have the designers, code writers and creators imbued their work (assumedly unintentionally) with their prejudices and biases, and to what extent has that made the systemic racism and privilege more entrenched.  And what damage has already been done, and what can we do to change the reality.  And that is unquestionably a Herculean challenge given that little of all the software is specific only to us.  But where it is at least semi-specific to us - such as perhaps in some grant managing software - can we identify where the coding or algorithms may be reflective of bias and prejudice?  Is that even possible?

With some exceptions, most programmers, coders, program / software creators are probably white men of a certain age.  To the extent we use what they have created in untold numbers of ways in our business and personal lives, continued exposure to their mind set biases, eventually has an impact on our thinking and our own experiences.  The more time we spend absorbing the thinking of others - particularly without any avenue of exchange about that thinking - the more it is likely to color our own thinking.  And most of that coloring goes on unnoticed.

A good illustration came up in another article this week - this one in the Nation - about Thought Leaders, in which the author quoted a passage from Barack Obama's (2006) book The Audacity of Hope:

 "Increasingly I found myself spending time with people of means—law firm partners and investment bankers, hedge fund managers and venture capitalists. As a rule, they were smart, interesting people, knowledgeable about public policy, liberal in their politics, expecting nothing more than a hearing of their opinions in exchange for their checks. But they reflected, almost uniformly, the perspectives of their class: the top 1 percent or so of the income scale that can afford to write a $2,000 check to a political candidate. They believed in the free market and an educational meritocracy; they found it hard to imagine that there might be any social ill that could not be cured by a high SAT score. They had no patience with protectionism, found unions troublesome, and were not particularly sympathetic to those whose lives were upended by the movements of global capital. Most were adamantly prochoice and antigun and were vaguely suspicious of deep religious sentiment.
[A]s a consequence of my fund-raising I became more like the wealthy donors I met, in the very particular sense that I spent more and more of my time above the fray, outside the world of immediate hunger, disappointment, fear, irrationality, and frequent hardship of the other 99 percent of the population—that is, the people that I’d entered public life to serve.”


We have all been spending an inordinate amount of time with those who have written the software and programs we have all been using in our computers for decades - including those people's biases and prejudices from their own specific upbringing and experiences.  They likely didn't know their baggage was included in their work, and we likely didn't know it either.  But in all probability that is the reality.  This is the ugly, insidious side of systemic, structural prejudice and privilege.

Fighting against this will be a lot harder than anyone could have possibly imagined, but the battles absolutely must be on all these deep levels - and especially for the future.  A biased, prejudiced AI landscape is beyond frightening.  It may pose a threat we are simply incapable of countering.

Have a good week.

Don't Quit.
Barry






Sunday, April 9, 2017

An Opportunity for the Arts (Maybe) as Retail Finds Itself in Deep Trouble

Good morning.
"And the beat goes on................."

The American retail industry continues to take big hits.  Increasingly, shoppers are moving online.  Not everywhere or for everything - yet, - but more and more the trend is away from bricks and mortar shops.  Maybe it's the ease and convenience - particularly in a world where we seem to all have less and less time to do the work that needs to get done.  And while shoppers still like to hit the stores when they can - especially for certain items like apparel stores are closing, downtowns, suburban areas and even malls have more and more shuttered, empty spaces.  Rents are coming down, as more retails are either in bankruptcy or nearing it.  And the retailers being hit range from the hip and trendy to the old stalwarts.

According to Bloomberg:

"The rapid descent of so many retailers has left shopping malls with hundreds of slots to fill, and the pain could be just beginning. More than 10 percent of U.S. retail space, or nearly 1 billion square feet, may need to be closed, converted to other uses or renegotiated for lower rent in coming years, according to data provided to Bloomberg by CoStar Group."

Consider this:

"Urban Outfitters Chief Executive Officer Richard Hayne didn’t mince words when he sized up the situation last month. Malls added way too many stores in recent years -- and way too many of them sell the same thing: apparel.
“This created a bubble, and like housing, that bubble has now burst,” he said. “We are seeing the results: Doors shuttering and rents retreating. This trend will continue for the foreseeable future and may even accelerate.”
Year-to-date store closings are already outpacing those of 2008, when the last U.S. recession was raging, according to Credit Suisse Group AG analyst Christian Buss. About 2,880 have been announced so far this year, compared with 1,153 for this period of 2016, he said in a report.

And not surprisingly, Amazon is far and away the catalyst of the trend:

"Even brands moving aggressively online have struggled to match the growth of market leader Amazon.com Inc.
The Seattle-based company accounted for 53 percent of e-commerce sales growth last year, with the rest of the industry sharing the remaining 47 percent, according to EMarketer Inc."

So what has that got to do with the arts?

It might be an opportunity for us.

We might explore a pilot program to see if the arts might do the same thing for malls and vacant retail space, whether downtown or in suburbia, that we have succeeded in doing for the revitalization and reinvention of some downtown areas across the country.  Arts organizations and artists might find a way to negotiate cheap rentals - perhaps supported by local government programs - for their being an attraction for people to come to the remaining retail space neighbors.  Artists and organizations looking for affordable space in an increasingly expensive real estate landscape might succeed in this situation where they bring with them multiple ways that might attract shoppers - from exhibitions to performances, to a vibrant arts ecosystem that is attractive to the public.

To go shopping and see artists at work, perhaps talk to them, watch rehearsals, maybe see performances, interact with arts education programs, poetry slams, dance companies, film makers and on and on might be a very attractive lure to the public.  And that might help retailers. And this might be a golden opportunity for us to target Millennials, even younger people, and to build public will in support of the arts.

It might be possible to negotiate some support for this kind of an effort from some of the major retail brands who are being threatened by the growth of Amazon and the online shopping presence.

And as more artists and arts organizations might occupy some of this space, in some instances it might grow into the spaces becoming de facto cultural centers.

It won't work everywhere, but it might work in some areas.

It's just a thought, but it seems to me our artists and our organizations need affordable work spaces, and venues to sell their product; we need new and expanded audiences; we need to interact with the public more directly beyond our normal channels,;the retail industry needs help in reinventing bricks and mortar shopping so it will attract shoppers; and the commercial real estate industry needs occupants and to stem the tide of the closures.  Win / Win?  Maybe.

I think this may be a possible opportunity for us to investigate and explore and I hope some funders will seed a couple of pilot programs along these lines to see if it might be something of benefit to us.

Have a good week.

Don't Quit
Barry






Sunday, April 2, 2017

Interviewing For New Hires

Good morning.
"And the beat goes on.........................."

Our organizations are arguably no better than our people who work for them.  Even for the largest of our organizations (which really aren't that large), every employee, every staff member, is crucial for the organization to operate at its optimum level.  And even if employees operate in silos, disconnected for the most part from each other, the organization's ecosystem is still a sum of its parts, for the work itself is connected.

Because our organizations are small, vertical promotion is often difficult, if not impossible.  Today, it is much more common for younger people to naturally expect to have multiple jobs at an ever earlier stage of their careers.  Limited and overtaxed budgets and problematic fundraising mean our sector, for the most part, and particularly at middle level management positions, can't always provide the level of compensation available in the private sector. There is the growing desire to run your own shop.  Long hours, frustration, and a variety of other forces are at work to make employee churn commonplace.   And so while we try to recruit the best people we can, and retain their services over time for continuity and other advantages, turnover is inevitable.

Finding the right people for open positions in a highly competitive job market is critical to our successes as organizations.

Once an organization narrows its candidates for an open position, reviews their resumes, and vets their recommendations and past performances, we invariably come down to two or three finalists. At that point the last stage of the hiring process is the in-person interview, where we try to glean information so we can make the best choice.

It is with that interview that I have some problems.  Increasingly, the interview has become some contest to see how clever we can be in designing the questions we ask.Too often now those questions don't really elicit the kinds of information that allow us to make intelligent, let alone, the best choice between candidates.  Too often, our interviews ignore what should rationally be our goals in favor of questions which put the interviewee on the spot - our thinking being that that will give us insight as to how the candidate will perform in our environment.

Questions such as: "What is your greatest weakness"; "How did you deal with failure?"; "Define your work ethic" - all sound reasonable, but suffer, I think, from stemming mostly from our attempts to be seen as wise and smart, and which simply don't tell us what we really need to know.  On one level, such questions are invasive and invite the interviewee to simply parrot back what they think we want to hear (which practice, I accept, is applicable to almost anything we ask, and now so widespread as to be expected - and that reality is yet another reason these kinds of questions simply mask the information that would be most valuable to us.)  To the extent we are trying to "game" the process with clever questions, the candidates will likewise try to game the process with answers they think fit our line of questioning.  We don't want the interview to be a contest of gaming each other.  We want it to be a frank, candid interchange between us; honest, transparent and fair to all.

Our obsession with everybody in the entire field needing to be a leader; our preoccupation with educational benchmarks in the form of degrees, which we equate with automatically being able to do the best job); and our laser like focus on where an applicant worked before - all color our thinking when we determine what we should ask of our finalists.

There are really only two major pieces of information we need to make an informed decision:

1) Can the applicant to a good (great) job in the position.  Do they have the experience, the thought processes, the vision and discipline to work at the organization and excel at the responsibilities that will be theirs.  How would they handle a specific challenge facing the person who will get the job.  I would be less interested in their weaknesses, or their failures, and more interested in their strengths - and I think it is the interviewer's job to determine their strengths by finding out what they would specifically do given a specific challenge.   More important than what they did in the past, is what they can do in the future - not in general, but for your organization.

2)  Every organization has its own culture.  Some are hierarchical; some authoritarian; some loose and flexible; some favor innovation and independent thinking and questioning and some have narrowly prescribed areas of decision making and how things are to be done.  The first thing that needs to be done is a realistic assessment of the organization's work culture - so that you can craft questions that will give you an idea whether or not the applicant will fit in.  The chemistry between the new hire and the extant work staff - and the organization itself - is arguably as important as their experience, expertise and vision.  If it ends up being a bad fit, the cost will be high to both the new hire and the current staff.  Too often, how the relationships might manifest get short changed or ignored in the interview process.  That is a big mistake.  One question you ought to ask yourself at the end of an interview is:  "Do I like this person?" because that is important.

It's easy to go online and seek some sample questions to ask prospective applicants in an interview, or for those that can afford to hire a search firm, to demur to that firm to come up with the questions.  But that is risky, for too often the questions then asked are the latest in the changing trend of what is fashionable at the time.  Generic questions may, if you are very lucky, give you some information that will actually be helpful in making your hiring decision, but don't count on it.  And it is an abrogation of the responsibility to take control of the process.  Nobody - certainly not a search firm - knows your organization like the people who work there.  Most search firms never bother to really research a client to determine what the ideal candidate would look like; rather they have a standardized "ideal' candidate profile that is basically a description of a candidate that is too perfect to exist in reality.

Some standardized questions might work for you, if you tweak them to fit your organization and the job slot to be filled.  But many more favored questions don't yield the kind of information they claim to.  Yes, you want to know how the candidate sees both work itself and the environment in which work is performed, and yes you want to know what the candidate values in relationships, their past successes and how they dealt with adversity - but be careful that the questions will actually elicit the information you need.  Too many questions elicit stock answers that are rehearsed and stray far from the facts.

Sometimes, it isn't what you ask of the candidate that tells you what you need to know. Sometimes, when you invite the candidate to ask you questions about the job, about the organization, those questions are more telling.  If the applicant doesn't have any questions for you, then they likely haven't done much research about your organization.  Ideally, the job interview isn't one sided; it's a conversation about the job, the applicant and the organization.  To get to that point you need to shift the power dynamic in your favor as the one doing the hiring to a more equal footing whereby the questions are back and forth and the interview becomes a discussion during which you can actually learn something about the applicant.

Hiring is often a crap shoot anyway.  You make the best decision you can given the information you have.  Sometimes you make a great hire.  Sometimes it turns out wrong and the relationship doesn't work.

But those hiring decisions, even at the lowest level of employees, are critically important and you need to treat them as such by thinking through the process before it starts.  That is true whether a Department Head, an Executive Director or the Board is making the hire.  The better you can assess the fit, the more likely you will make a wise hire.

So, please spend some time drafting questions that will give you information about your job candidates as to how they will fit into your organization, and how they will handle the actual job they will be expected to perform.  You need to be as honest with yourself in preparing for the interview as you hope the applicant will be in responding to your inquiries.  THINK about it.  Knowing everything you do about your organization and the job, if you were the candidate, what questions ought to be asked to determine if you would be the right person to hire.  My guess is your biggest failure or other such irrelevant and invasive questions wouldn't be on your list.

Good luck.

Have a great week.

Don't Quit
Barry