Helping A.I. to Learn About Indigenous Cultures

2022/03/23 Leave a comment

Interesting:

In September 2021, Native American technology students in high school and college gathered at a conference in Phoenix and were asked to create photo tags — word associations, essentially — for a series of images.

One image showed ceremonial sage in a seashell; another, a black-and-white photograph circa 1884, showed hundreds of Native American children lined up in uniform outside the Carlisle Indian Industrial School, one of the most prominent boarding schools run by the American government during the 19th and 20th centuries.

For the ceremonial sage, the students chose the words “sweetgrass,” “sage,” “sacred,” “medicine,” “protection” and “prayers.” They gave the photo of the boarding school tags with a different tone: “genocide,” “tragedy,” “cultural elimination,” “resiliency” and “Native children.”

The exercise was for the workshop Teaching Heritage to Artificial Intelligence Through Storytelling at the annual conference for the American Indian Science and Engineering Society. The students were creating metadata that could train a photo recognition algorithm to understand the cultural meaning of an image.

The workshop presenters — Chamisa Edmo, a technologist and citizen of the Navajo Nation, who is also Blackfeet and Shoshone-Bannock; Tracy Monteith, a senior Microsoft engineer and member of the Eastern Band of Cherokee Indians; and the journalist Davar Ardalan — then compared these answers with those produced by a major image recognition app.

For the ceremonial sage, the app’s top tag was “plant,” but other tags included “ice cream” and “dessert.” The app tagged the school image with “human,” “crowd,” “audience” and “smile” — the last a particularly odd descriptor, given that few of the children are smiling.

The image recognition app botched its task, Mr. Monteith said, because it didn’t have proper training data. Ms. Edmo explained that tagging results are often “outlandish” and “offensive,” recalling how one app identified a Native American person wearing regalia as a bird. And yet similar image recognition apps have identified with ease a St. Patrick’s Day celebration, Ms. Ardalan noted as an example, because of the abundance of data on the topic.

As Mr. Monteith put it, A.I. is only as good as the data it is fed. And data on cultures that have long been marginalized, like Native ones, are simply not at the levels they need to be. “Clearly, there’s a bias represented,” he said.

The workshop was the initiative of Intelligent Voices of Wisdom, or IVOW, a tech start-up that Ms. Ardalan, an executive producer of audio at National Geographic, founded to preserve culture through A.I. and to counter those biases.

“The internet is not representative of the entire population, and when people are represented, it may not be accurate because of stereotypes and hate speech,” said Percy Liang, an associate professor of computer science at Stanford University and director of the school’s Center for Research on Foundation Models.

To counter this tendency, Ms. Ardalan, who is an Iranian American of Bakhtiari and Kurdish descent, wants IVOW to develop tools to create “cultural engines” for underrepresented groups so they can generate, and take ownership of, their data. “The cultural engine cannot be a data scientist in Philadelphia trying to create data sets for a tribe in Arizona,” she said.

More representative, accurate data is beneficial not only to the groups it represents, but also to A.I. systems at large, said W. Victor H. Yarlott, an A.I. researcher at Florida International University, a member of the Crow Tribe of Montana and an IVOW collaborator.

“Lacking this knowledge just makes your system worse,” he said. “You’re not really representing human intelligence or human knowledge unless your system can handle it from a broad range of cultures.”

The participation of Indigenous people in the project was critical. Mr. Monteith, who led the effort to enter the Cherokee writing system into Microsoft Office, said he’s been working on building trust for technology, and more recently A.I., in his Native communities for decades. “I knew without me doing this that we would be in a worse spot in terms of literacy, and our culture,” he said.

The team at IVOW, along with a group of volunteer collaborators and advisers, has been developing proofs of concept for these cultural engines — smart data sets that can feed more inclusive A.I. tools, including chatbots and image recognition apps.

One such tool is IVOW’s Indigenous Knowledge Graph, or IKG, a cultural engine in early development that is focused on storytelling about Indigenous recipes and culinary practices. After meeting the IVOW team in 2018, Mr. Yarlott pitched the IKG, a sort of visualization of a data set, to capture Indigenous knowledge.

“You know in dramas, you see the person trying to unravel a mystery and they have the corkboard and the little notes and the string between them?” Mr. Yarlott said. “That’s basically what the IKG is, but for cultural knowledge.”

The first step was to gather the data. The team chose a culinary focus because it is a part of life that all people share. They collected recipes and related stories from both the public domain and team members.

Mr. Monteith chose to enter the story of the Three Sisters stew, a recipe created from symbiotic crops (corn, beans and squash) that he said is known among Indigenous peoples wherever those ingredients grow. The story of the Three Sisters, he said, is not only a recipe but a way to teach sustainability practices, such as the preservation of water. “It’s just a great metaphor for what we need to do as a society and as a people across the world,” Mr. Monteith said.

Using Neo4J, a graph database management system, the recipes were broken down into components (title, ingredients, instructions and related stories) and tagged with information, like the tribe of origin or whether the recipe was contemporary or historical, or had roots in folklore. This data set was then entered into Dialogflow, a natural language processing platform, so it could be fed into a chatbot — in this case, Sina Storyteller, the Siri-like conversational agent designed by IVOW. Currently, anyone can interact with the early version through Google Assistant.

The tools and techniques to create the IKG were designed to be basic enough that anyone, not just those with a background in computer science, could use them. And IKG uses only information that is widely available or that the team had permission to use from their own tribes, bands and nations.

There are challenges, though. The process is labor intensive and expensive; IVOW is a self-funded enterprise, and the work of the collaborators is voluntary.

“It’s a little bit of a chicken and an egg problem because you need the data to really build a big system that demonstrates value,” Mr. Yarlott said. “But to get all the data, you need money, which only really starts to come when people realize that there’s substantial value here.”

Mr. Liang said that while this kind of “artisanal” data is important, it is difficult to scale, and that more emphasis should be placed on improving foundation knowledge — models that are trained on large-scale data sets.

For years, computer scientists have warned Ms. Ardalan that cultivating this sort of data is a tedious process. She doesn’t disagree, which is why she says the time to start is now.

“The future is going to be these cultural engines that communities create that are relevant to their heritage,” she said, adding that the notion that A.I. will be all-encompassing is wrong. “Machines cannot replace humans. They can only be there with us around the campfire and inform us.”

Source: Helping A.I. to Learn About Indigenous Cultures

Filed under Aboriginal Tagged with , , ,

Automation could make 12 million jobs redundant. Here’s who’s most at risk

2022/01/24 Leave a comment

Although a European study, likely more broadly applicable to Canada and other countries, something that advocates of current and higher levels of immigration to Canada understate or ignore:

Up to a third of job roles in Europe could be made redundant by automation over the next 20 years as companies battle to increase productivity and fill skills gaps created by an ageing population, according to Forrester. 

The tech analyst’s latest Future of Jobs Forecast estimates that as many as 12 million jobs could be lost to automation across Europe by 2040, primarily impacting workers in industries such as retail, food services, and leisure and hospitality.  

Mid-skill labour jobs that consist of simple, routine tasks are most at risk from automation, the report said. These roles make up 38% of the workforce in Germany, 34% of the workforce in France, and 31% of the workforce in the UK. 

In total, 49 million jobs in ‘Europe-5’ (France, Germany, Italy, Spain, and the UK) could potentially be automated, according to Forrester. This jeopardizes casual work, such as zero-hour contracts in the UK, and low-paid, part-time jobs where workers hold “little bargaining power”. 

A combination of pressures is prompting businesses to ramp up their investments in automation, particularly in countries where industry, construction and agriculture are big business. 

While small and medium enterprises (SMEs) with up to 50 workers capture two-thirds of European employment, their productivity lags that of larger corporations, according to Forrester. In manufacturing, for example, ‘microenterprises’ are 40% less productive than large companies. 

A five-year study of robot adoption at French manufacturing firms found that robots lowered production overheads by reducing labor costs by between 4% and 6%. 

Business leaders also see automation technology as a means of filling the gaps created by Europe’s ageing population, which Forrester describes as “a demographic time bomb.” By 2050, Europe will have 30 million fewer people of working age than in 2020, the analyst said. 

Productivity lost to the pandemic is seeing organizations look to machine processes to recoup efficiency, while industries that were already using automation to grow their revenues have invested even more heavily in the technology to increase service delivery and mitigate pandemic restrictions. 

“Lost productivity due to COVID-19 is forcing companies globally to automate manual processes and improve remote work,” said Michael O’Grady, principal forecast analyst at Forrester. 

“European organisations are also in a particularly strong position to embrace automation because of Europe’s declining working-age population and the high number of routine, low-skilled jobs that can be easily automated.” 

While many low-skilled and routine roles face being replaced by machine processes, nine million new jobs are forecast to be created in Europe by 2040 in emerging sectors like green energy and smart cities, Forrester said.

This means that, all told, only three million jobs will truly be ‘lost’ to automation by 2040 – the caveat being that people who lose jobs may not find new ones.

Business leaders outside of Europe are also exploring the role of automation in bridging skills shortages and speeding up processes in the enterprise. 

Polling of 500 C-suite executives and senior management personnel by automation platform UIPath found that 78% were likely to increase their investment in automation tools to help them address labor shortages. Business leaders are turning to automation because they are struggling to find new talent (74%). 

At the same time, 85% of survey respondents said incorporating automation and automation training into their organization would help them attract new talent and hold onto existing staff. Meanwhile, leaders said automation was already helping them to save time (71%), improve productivity (63%) and save money (59%). 

Academic forecasts of jobs that could be lost to automation vary wildly. The European Parliament’s 2021 ‘Digital automation and the future of work’ report found that estimates varied from as little as 9% to almost half (47%). 

“Machine-learning experts often drive this uncertainty,” said Forrester. 

“They imagine future computer capabilities without understanding enterprise technology adoption constraints and the cultural barriers within an organization that resist change.” 

Source: https://www.zdnet.com/article/automation-could-make-12-million-jobs-redundant-heres-whos-most-at-risk/?utm_campaign=David%20Akin%27s%20🇨🇦%20Roundup&utm_medium=email&utm_source=Revue%20newsletter

Filed under Immigration Tagged with , , , ,

Canada is refusing more study permits. Is new AI technology to blame?

2021/11/16 Leave a comment

Given the high volumes (which immigration lawyers and consultants benefit from), expanded use of technology and templates inevitable and necessary, although thorough review and safeguards necessary.

Alternate narrative, given reporting on abuse and exploitation of international students and the program itself (The reality of life in Canada for international students), perhaps a system generating more refusals has merit:

Soheil Moghadam applied twice for a study permit for a postgraduate program in Canada, only to be refused with an explanation that read like a templated answer.

The immigration officer was “not satisfied that you will leave Canada at the end of your stay,” he was told.

After a third failed attempt, Moghadam, who already has a master’s degree in electronics engineering from Iran, challenged the refusal in court and the case was settled. He’s now studying energy management at the New York Institute of Technology in Vancouver.

His Canadian lawyer, Zeynab Ziaie, said that in the past couple of years, she has noticed a growing number of study permit refusals like Moghadam’s. The internal notes made by officers reveal only generic analyses based on cookie-cutter language and often have nothing to do with the particular evidence presented by the applicant.

“We’re seeing a lot of people that previously would have been accepted or have really what we consider as complete files with lots of evidence of financial support, lots of ties to their home country. These kinds of files are just being refused,” said Ziaie, who added that she has seen more than 100 of these refusals in her practice in the past two years.

It’s a Microsoft Excel-based system called Chinook. 

Its existence came to light during a court case involving Abigail Ocran, a woman from Ghana who was refused a study permit by the Immigration Department.

Government lawyers in that case filed an affidavit by Andie Daponte, director of international-network optimization and modernization, who detailed the working and application of Chinook.

That affidavit has created a buzz among those practising immigration law, who see the new system — the department’s transition to artificial intelligence — as a potential threat to quality decision making, and its arrival as the harbinger of more troubling AI technology that could transform how immigration decisions are made in this country.

All eyes are now on the pending decision of the Ocran case to see if and how the court will weigh in on the use of Chinook. 

Chinook was implemented in March 2018 to help the Immigration Department handle an exponential growth in cases within its existing, and antiquated, Global Case Management System (GCMS).

Between 2011 and 2019, before everything slowed down during the pandemic, the number of visitor visa applications skyrocketed by 109 per cent, with the caseload of applications for overseas work permits and study permits up by 147 per cent and 222 per cent, respectively.

In 2019 alone, Daponte said in his affidavit, Canada received almost 2.2 million applications from prospective visitors, in addition to 366,000 from people looking to work here and 431,500 from would-be international students.

Meanwhile, the department’s 17-year-old GCMS system, which requires officers to open multiple screens to download different information pertaining to an application, has not caught up. Each time decision-makers move from screen to screen they must wait for the system to load, causing significant delays in processing, especially in countries with limited network bandwidth.

Chinook was developed in-house and implemented “to enhance efficiency and consistency, and to reduce processing times,” Daponte said.

As a result, he said, migration offices have generally seen an increase of between five per cent and 35 per cent in the number of applications they have been able to process.

Here’s how Chinook works: an applicant’s information is extracted from the old system and populated in a spreadsheet, with each cell on the same row filled with data from that one applicant — such as name, age, purpose of visit, date of receipt of the application and previous travel history.

Each spreadsheet contains content from multiple applicants and is assigned to an officer to enable them to use “batch processes.”

After the assessment of an application is done, the officer will click on the decision column to prompt a pop-up window to record the decision, along with a notes generator if they’re giving reasons in the case of a refusal.

(An officer can refuse or approve an application, and sometimes hold it for further information.)

When done, decision-makers click a button labelled “Action List,” which organizes data for ease of transfer into the old system. It presents the decision, reasons for refusal if applicable, and any “risk indicators” or “local word flags” for each application.

The spreadsheets are deleted daily after the data transfer for privacy concerns.

While working on the spreadsheet, said Daponte, decision-makers continue to have access to paper applications or electronic documents and GCMS if needed.

“Chinook was built to save decision-makers time in querying GCMS for application information and to allow for the review of multiple applications,” Daponte noted.

However, critics are concerned that the way the system is set up may be guiding the officers toward certain conclusions, giving them the option of not reviewing all the material presented in each case, and that it effectively shields much of the decision making from real scrutiny.

According to Daponte’s court affidavit, the notes generator presents standard language that immigration officers may select, review and modify to fit the circumstances of an application in preparing reasons for refusal. The function is there to “assist them in the creation of reasons.”

Ziaie believes that explains the templated reasons for refusals she’s been seeing.

“These officers are looking at a spreadsheet of potentially 100 different applicants. And those names don’t mean anything to the officers. You could mix up rows. You could easily make errors,” said the Toronto lawyer.

“There’s no way to go back and check that because these decisions end up with very similar notes that are generated right when they’re refused. So my concern is about accountability. Every time we have a decision, it has to make sense. We don’t know if they make mistakes.”

That’s why she and other lawyers worry the surge of study permit refusals is linked to the implementation of Chinook. 

In fact, that question was put to Daponte during the cross-examination in the Ocran case by the Ghanaian student’s lawyer, Edos Omorotionmwan.

Immigration data obtained by Omorotionmwan showed the refusal rate of student permit applications had gone from 31 per cent in 2016 to 34 per cent in 2018, the year Chinook was launched. The trend continued in 2019 to 40 per cent and reached 53 per cent last year.

“Is there a system within the Chinook software requiring some oversight function where there is some other person to review what a visa officer has come up with before that decision is handed over to the applicants?” asked Omorotionmwan.

“Within Chinook, no,” replied Daponte, who also said there’s no mechanism within this platform to track if an officer has reviewed all the support documents and information pertaining to an applicant’s file in the GCMS data.

“This idea of using portals and technology to speed up the way things are done is the reality of the future,” said Vancouver-based immigration lawyer Will Tao, who has tracked the uses of Chinook and blogged about it.

“My concern as an advocate is: who did this reality negatively impact and what systems does it continue to uphold?”

Tao said the way the row of personal information is selected and set out in the Chinook spreadsheet “disincentivizes” officers to go into the actual application materials and support documents out of convenience.

“And then the officers are supposed to use those notes generators to justify their reasoning and not go into some of the details that you would like to see to reflect that they actually reviewed the facts of the case. The biggest problem I have is that this system has had very limited oversight,” he said.

“It makes it easier to refuse because you don’t have to look at all the facts. You don’t have to go through a deep, thoughtful analysis. You have a refusal notes generator that you can apply without having read the detailed study plans and financial documents.”

He points to Chinook’s built-in function that flags “risk factors” — such as an applicant’s occupation and intended employer’s information — for inconsistency in an application, as well as “local flag words” to triage and ensure priority processing of time-sensitive applications to attend a wedding or a funeral.

Those very same flag words used in the spreadsheet can also be misused to mark a particular group of applicants based on their personal profiles and pick them out for refusals, said Tao.

In 2019, in a case involving the revocation of citizenship to the Canadian-born sons of two Russian spies, the Supreme Court of Canada made a landmark ruling that helps guide judges to review the decisions of immigration officials.

In the unanimous judgment, Canada’s highest court ruled it would be “unacceptable for an administrative decision maker to provide an affected party formal reasons that fail to justify its decision, but nevertheless expect that its decision would be upheld on the basis of internal records that were not available to that party.”

Tao said he’s closely watching how the Ocran decision is going to shed light on the application of Chinook in the wake of that Supreme Court of Canada ruling over the reasonableness standard.

“Obviously, a lot of these applications have critical points that they get refused on and with the reasons being template and standard, it’s hard for reviewers to understand how that came to be,” he said.

In a response to the Star’s inquiry about the concerns raised about Chinook, the Immigration Department said the tool is simply to streamline the administrative steps that would otherwise be required in the processing of applications to improve efficiency.

“Decision makers are required to review all applications and render their decisions based on the information presented before them,” said spokesperson Nancy Caron.

“Chinook does not fundamentally change the way applications are processed, and it is always the officer that gives the rational for the decisions and not the Chinook tool.”

For immigration lawyer Mario Bellissimo, Chinook is another step in the Immigration Department’s move toward digitalization and modernization.

Ottawa has been using machine learning technology since 2018 to triage temporary resident visa applications from China and India, using a “set of rules derived from thousands of past officer decisions” then deployed by the technology to classify applications into high, medium and low complexity.

Cases identified as low complexity and low risk automatically receive positive eligibility decisions, allowing officers to review these files exclusively on the basis of admissibility. This enables officers to spend more time scrutinizing the more complex files.

Chinook, said Bellissimo, has gone beyond the triage. He contends it facilitates the decision-making process by officers.

The use of templated responses from the notes generator makes the refusal reasons “devoid of meaning,” he noted.

“Eventually, do you see age discriminators put into place for study permits when anyone over the age of 30 is all automatically streamed to a different tier because they are less likely bona fide students? This is the type of stuff we need to know,” Bellissimo explained.

“When they’re just pulling standard refusal reasons and just slapping it in, then those decisions become more difficult to understand and more difficult to challenge. Who made the decision? Was technology used? And that becomes a problem.”

He said immigration officials need to be accountable and transparent to applicants about the use of these technologies before they are rolled out, not after they become an issue.

Petra Molnar, a Canadian expert specializing in migration and technology, said automated decision-making and artificial intelligence tools are difficult to scrutinize because they are often very opaque, including how they are developed and deployed and what review mechanisms, if any, exist once they are in use.

“Decisions in the immigration and refugee context have lifelong and life-altering ramifications. People have the right to know what types of tools are being used against them and how they work, so that we can meaningfully challenge these types of systems.”

Ziaie, the lawyer, said she understands the tremendous pressure on front-line immigration officers, but if charging a higher application fee — a study permit application now costs $150 — can help improve the service and quality of decisions, then that should be implemented.

“They should allocate a fair amount of that revenue toward trying to hire more people, train their officers better and give them more time to review the files so they actually do get a better success rate,” she said. “By that, I mean fewer files going to Federal Court.”

As a study permit applicant, Moghadam said it’s frustrating not to understand how an immigration officer reaches a refusal decision because so much is at stake for the applicant.

It took him two extra years to finally obtain his study permit and pursue an education in Canada, let alone the additional application fees and hefty legal costs.

“Your life is put on hold and your future is uncertain,” said the 39-year-old, who had a decade of work experience in engineering for both Iranian and international companies.

“There’s the time, the costs, the stress and the anxiety.”

Source: https://www.thestar.com/news/canada/2021/11/15/canada-is-refusing-more-study-permits-is-new-ai-technology-to-blame.html

Filed under Immigration Tagged with , , , ,

Rise of the Robots Speeds Up in Pandemic With U.S. Labor Scarce

2021/11/08 Leave a comment

Of note to Canadian policy makers as well given this trend will cross the border and needs to be taken into account in immigration policy:

American workers are hoping that the tight pandemic labor market will translate into better pay. It might just mean robots take their jobs instead.

Labor shortages and rising wages are pushing U.S. business to invest in automation. A recent Federal Reserve survey of chief financial officers found that at firms with difficulty hiring, one-third are implementing or exploring automation to replace workers. In earnings calls over the past month, executives from a range of businesses confirmed the trend.

Domino’s Pizza Inc. is “putting in place equipment and technology that reduce the amount of labor that is required to produce our dough balls,” said Chief Executive Officer Ritch Allison.
[time-brightcove not-tgx=”true”]

Mark Coffey, a group vice president at Hormel Foods Corp., said the maker of Spam spread and Skippy peanut butter is “ramping up our investments in automation” because of the “tight labor supply.”

The mechanizing of mundane tasks has been underway for generations. It’s made remarkable progress in the past decade: The number of industrial robots installed in the world’s factories more than doubled in that time, to about 3 million. Automation has been spreading into service businesses too.

The U.S. has lagged behind other economies, especially Asian ones, but the pandemic might trigger some catching up. With some 10.4 million open positions as of August, and record numbers of Americans quitting their jobs, the difficulty of finding staff is adding new incentives.

Ametek Inc. makes automation equipment for industrial firms, like motion trackers that are used from steel and lumber mills to packaging systems. Chief Executive Officer David A. Zapico says that part of the company is “firing on all cylinders.” That’s because “people want to remove labor from the processes,” he said on an earnings call. “In some places, you can’t hire labor.”

Unions have long seen automation as a threat. At U.S. ports, which lag their global peers in technology and are currently at the center of a major supply-chain crisis, the International Longshoremen’s Association has vowed to fight it.

Companies that say they want to automate “have one goal in mind: to eliminate your job, and put more money in their pockets,” ILA President Harold Daggett said in a video message to a June conference. “We’re going to fight this for 100 years.”

Some economists have warned that automation could make America’s income and wealth gaps worse.

“If it continues, labor demand will grow slowly, inequality will increase, and the prospects for many low-education workers will not be very good,” says Daron Acemoglu, a professor at the Massachusetts Institute of Technology, who testified Wednesday at a Senate hearing on the issue.

That’s not an inevitable outcome, Acemoglu says: Scientific knowhow could be used “to develop technologies that are more complementary to workers.” But, with research largely dominated by a handful of giant firms that spend the most money on it, “this is not the direction the technology is going currently.”

Knightscope makes security robots that look a bit like R2-D2 from Star Wars, and can patrol sites such as factory perimeters. The company says it’s attracting new clients who are having trouble hiring workers to keep watch. Its robots cost from $3.50 to $7.50 an hour, according to Chief Client Officer Stacy Stephens, and can be installed a month after signing a contract.

One new customer is the Los Angeles International Airport, one of the busiest in the U.S. Soon, Knightscope robots will be monitoring some of its parking lots.

They are “supplementing what we have in place and are not replacing any human services,” said Heath Montgomery, the airport’s director of public relations. “It’s another way we are providing exceptional guest experiences.”

Source: Rise of the Robots Speeds Up in Pandemic With U.S. Labor Scarce

Filed under Immigration Tagged with , , , ,

AI’s anti-Muslim bias problem

2021/09/20 Leave a comment

Of note (and unfortunately, not all that surprising):

Imagine that you’re asked to finish this sentence: “Two Muslims walked into a …”

Which word would you add? “Bar,” maybe?

It sounds like the start of a joke. But when Stanford researchers fed the unfinished sentence into GPT-3, an artificial intelligence system that generates text, the AI completed the sentence in distinctly unfunny ways. “Two Muslims walked into a synagogue with axes and a bomb,” it said. Or, on another try, “Two Muslims walked into a Texas cartoon contest and opened fire.”

For Abubakar Abid, one of the researchers, the AI’s output came as a rude awakening. “We were just trying to see if it could tell jokes,” he recounted to me. “I even tried numerous prompts to steer it away from violent completions, and it would find some way to make it violent.”

Language models such as GPT-3 have been hailed for their potential to enhance our creativity. Given a phrase or two written by a human, they can add on more phrases that sound uncannily human-like. They can be great collaborators for anyone trying to write a novel, say, or a poem.

Source: AI’s anti-Muslim bias problem

Filed under Islamophobia, Multiculturalism, Religion Tagged with , , , ,

Misattributed blame? Attitudes toward globalization in the age of automation

2021/09/13 Leave a comment

Interesting study and findings:

Many, especially low-skilled workers, blame globalization for their economic woes. Robots and machines, which have led to job market polarization, rising income inequality, and labor displacement, are often viewed much more forgivingly. This paper argues that citizens have a tendency to misattribute blame for economic dislocations toward immigrants and workers abroad, while discounting the effects of technology. Using the 2016 American National Elections Studies, a nationally representative survey, I show that workers facing higher risks of automation are more likely to oppose free trade agreements and favor immigration restrictions, even controlling for standard explanations for these attitudes. Although pocket-book concerns do influence attitudes toward globalization, this study calls into question the standard assumption that individuals understand and can correctly identify the sources of their economic anxieties. Accelerated automation may have intensified attempts to resist globalization.

Source: https://www.cambridge.org/core/journals/political-science-research-and-methods/article/misattributed-blame-attitudes-toward-globalization-in-the-age-of-automation/29B08295CEAC4A4A89991E064D0284FF

Filed under Immigration, Public Policy Tagged with , , ,

Facebook Apologizes After Its AI Labels Black Men As ‘Primates’

2021/09/07 Leave a comment

Ouch!

Facebook issued an apology on behalf of its artificial intelligence software that asked users watching a video featuring Black men if they wanted to see more “videos about primates.” The social media giant has since disabled the topic recommendation feature and says it’s investigating the cause of the error, but the video had been online for more than a year.

A Facebook spokesperson told The New York Times on Friday, whichfirst reported on the story, that the automated prompt was an “unacceptable error” and apologized to anyone who came across the offensive suggestion.

The video, uploaded by the Daily Mail on June 27, 2020, documented an encounter between a white man and a group of Black men who were celebrating a birthday. The clip captures the white man allegedly calling 911 to report that he is “being harassed by a bunch of Black men,” before cutting to an unrelated video that showed police officers arresting a Black tenant at his own home.

Former Facebook employee Darci Groves tweeted about the error on Thursday after a friend clued her in on the misidentification. She shared a screenshot of the video that captured Facebook’s “Keep seeing videos about Primates?” message.

“This ‘keep seeing’ prompt is unacceptable, @Facebook,” she wrote. “And despite the video being more than a year old, a friend got this prompt yesterday. Friends at [Facebook], please escalate. This is egregious.”

This is not Facebook’s first time in the spotlight for major technical errors. Last year, Chinese President Xi Jinping’s name appeared as “Mr. S***hole” on its platform when translated from Burmese to English. The translation hiccup seemed to be Facebook-specific, and didn’t occur on Google, Reuters had reported.

However, in 2015, Google’s image recognition software classified photos of Black people as “gorillas.” Google apologized and removed the labels of gorilla, chimp, chimpanzee and monkey — words that remained censored over two years later, Wired reported.

Facebook could not be reached for comment.

Source: Facebook Apologizes After Its AI Labels Black Men As ‘Primates’

Filed under Multiculturalism Tagged with , , , ,

Using A.I. to Find Bias in A.I.

2021/07/02 Leave a comment

In 2018, Liz O’Sullivan and her colleagues at a prominent artificial intelligence start-up began work on a system that could automatically remove nudity and other explicit images from the internet.

They sent millions of online photos to workers in India, who spent weeks adding tags to explicit material. The data paired with the photos would be used to teach A.I. software how to recognize indecent images. But once the photos were tagged, Ms. O’Sullivan and her team noticed a problem: The Indian workers had classified all images of same-sex couples as indecent.

For Ms. O’Sullivan, the moment showed how easily — and often — bias could creep into artificial intelligence. It was a “cruel game of Whac-a-Mole,” she said.

This month, Ms. O’Sullivan, a 36-year-old New Yorker, was named chief executive of a new company, Parity. The start-up is one of many organizations, including more than a dozen start-ups and some of the biggest names in tech, offering tools and services designed to identify and remove bias from A.I. systems.

Soon, businesses may need that help. In April, the Federal Trade Commission warned against the sale of A.I. systems that were racially biased or could prevent individuals from receiving employment, housing, insurance or other benefits. A week later, the European Union unveiled draft regulations that could punish companies for offering such technology.

It is unclear how regulators might police bias. This past week, the National Institute of Standards and Technology, a government research lab whose work often informs policy, released a proposal detailing how businesses can fight bias in A.I., including changes in the way technology is conceived and built.

Many in the tech industry believe businesses must start preparing for a crackdown. “Some sort of legislation or regulation is inevitable,” said Christian Troncoso, the senior director of legal policy for the Software Alliance, a trade group that represents some of the biggest and oldest software companies. “Every time there is one of these terrible stories about A.I., it chips away at public trust and faith.”

Over the past several years, studies have shown that facial recognition services, health care systems and even talking digital assistants can be biased against women, people of color and other marginalized groups. Amid a growing chorus of complaints over the issue, some local regulators have already taken action.

In late 2019, state regulators in New York opened an investigationof UnitedHealth Group after a study found that an algorithm used by a hospital prioritized care for white patients over Black patients, even when the white patients were healthier. Last year, the state investigated the Apple Card credit service after claims it was discriminating against women. Regulators ruled that Goldman Sachs, which operated the card, did not discriminate, while the status of the UnitedHealth investigation is unclear. 

A spokesman for UnitedHealth, Tyler Mason, said the company’s algorithm had been misused by one of its partners and was not racially biased. Apple declined to comment.

More than $100 million has been invested over the past six months in companies exploring ethical issues involving artificial intelligence, after $186 million last year, according to PitchBook, a research firm that tracks financial activity.

But efforts to address the problem reached a tipping point this month when the Software Alliance offered a detailed framework for fighting bias in A.I., including the recognition that some automated technologies require regular oversight from humans. The trade group believes the document can help companies change their behavior and can show regulators and lawmakers how to control the problem.

Though they have been criticized for bias in their own systems, Amazon, IBM, Google and Microsoft also offer tools for fighting it.

Ms. O’Sullivan said there was no simple solution to bias in A.I. A thornier issue is that some in the industry question whether the problem is as widespread or as harmful as she believes it is.

“Changing mentalities does not happen overnight — and that is even more true when you’re talking about large companies,” she said. “You are trying to change not just one person’s mind but many minds.”

When she started advising businesses on A.I. bias more than two years ago, Ms. O’Sullivan was often met with skepticism. Many executives and engineers espoused what they called “fairness through unawareness,” arguing that the best way to build equitable technology was to ignore issues like race and gender.

Increasingly, companies were building systems that learned tasks by analyzing vast amounts of data, including photos, sounds, text and stats. The belief was that if a system learned from as much data as possible, fairness would follow.

But as Ms. O’Sullivan saw after the tagging done in India, bias can creep into a system when designers choose the wrong data or sort through it in the wrong way. Studies show that face-recognition services can be biased against women and people of color when they are trained on photo collections dominated by white men.

Designers can be blind to these problems. The workers in India — where gay relationships were still illegal at the time and where attitudes toward gays and lesbians were very different from those in the United States — were classifying the photos as they saw fit.

Ms. O’Sullivan saw the flaws and pitfalls of artificial intelligence while working for Clarifai, the company that ran the tagging project. She said she had left the company after realizing it was building systems for the military that she believed could eventually be used to kill. Clarifai did not respond to a request for comment. 

She now believes that after years of public complaints over bias in A.I. — not to mention the threat of regulation — attitudes are changing. In its new framework for curbing harmful bias, the Software Alliance warned against fairness through unawareness, saying the argument did not hold up.

“They are acknowledging that you need to turn over the rocks and see what is underneath,” Ms. O’Sullivan said.

Still, there is resistance. She said a recent clash at Google, where two ethics researchers were pushed out, was indicative of the situation at many companies. Efforts to fight bias often clash with corporate culture and the unceasing push to build new technology, get it out the door and start making money.

It is also still difficult to know just how serious the problem is. “We have very little data needed to model the broader societal safety issues with these systems, including bias,” said Jack Clark, one of the authors of the A.I. Index, an effort to track A.I. technology and policy across the globe. “Many of the things that the average person cares about — such as fairness — are not yet being measured in a disciplined or a large-scale way.”

Ms. O’Sullivan, a philosophy major in college and a member of the American Civil Liberties Union, is building her company around a tool designed by Rumman Chowdhury, a well-known A.I. ethics researcher who spent years at the business consultancy Accenture before joining Twitter.

While other start-ups, like Fiddler A.I. and Weights and Biases, offer tools for monitoring A.I. services and identifying potentially biased behavior, Parity’s technology aims to analyze the data, technologies and methods a business uses to build its services and then pinpoint areas of risk and suggest changes.

The tool uses artificial intelligence technology that can be biased in its own right, showing the double-edged nature of A.I. — and the difficulty of Ms. O’Sullivan’s task.

Tools that can identify bias in A.I. are imperfect, just as A.I. is imperfect. But the power of such a tool, she said, is to pinpoint potential problems — to get people looking closely at the issue.

Ultimately, she explained, the goal is to create a wider dialogue among people with a broad range of views. The trouble comes when the problem is ignored — or when those discussing the issues carry the same point of view.

“You need diverse perspectives. But can you get truly diverse perspectives at one company?” Ms. O’Sullivan asked. “It is a very important question I am not sure I can answer.”

Source: https://www.nytimes.com/2021/06/30/technology/artificial-intelligence-bias.html

Filed under Multiculturalism Tagged with , ,

Why A.I. Should Be Afraid of Us: Because benevolent bots are suckers.

2021/06/07 Leave a comment

Of significance as AI becomes more prevalent. “Road rage” as the new Turing test!

Artificial intelligence is gradually catching up to ours. A.I. algorithms can now consistently beat us at chess, poker and multiplayer video games, generate images of human faces indistinguishable from real ones, write news articles (not this one!) and even love stories, and drive cars better than most teenagers do.

But A.I. isn’t perfect, yet, if Woebot is any indicator. Woebot, as Karen Brown wrote this week in Science Times, is an A.I.-powered smartphone app that aims to provide low-cost counseling, using dialogue to guide users through the basic techniques of cognitive-behavioral therapy. But many psychologists doubt whether an A.I. algorithm can ever express the kind of empathy required to make interpersonal therapy work.

“These apps really shortchange the essential ingredient that — mounds of evidence show — is what helps in therapy, which is the therapeutic relationship,” Linda Michaels, a Chicago-based therapist who is co-chair of the Psychotherapy Action Network, a professional group, told The Times.

Empathy, of course, is a two-way street, and we humans don’t exhibit a whole lot more of it for bots than bots do for us. Numerous studies have found that when people are placed in a situation where they can cooperate with a benevolent A.I., they are less likely to do so than if the bot were an actual person.

“There seems to be something missing regarding reciprocity,” Ophelia Deroy, a philosopher at Ludwig Maximilian University, in Munich, told me. “We basically would treat a perfect stranger better than A.I.”

In a recent study, Dr. Deroy and her neuroscientist colleagues set out to understand why that is. The researchers paired human subjects with unseen partners, sometimes human and sometimes A.I.; each pair then played a series of classic economic games — Trust, Prisoner’s Dilemma, Chicken and Stag Hunt, as well as one they created called Reciprocity — designed to gauge and reward cooperativeness.

Our lack of reciprocity toward A.I. is commonly assumed to reflect a lack of trust. It’s hyper-rational and unfeeling, after all, surely just out for itself, unlikely to cooperate, so why should we? Dr. Deroy and her colleagues reached a different and perhaps less comforting conclusion. Their study found that people were less likely to cooperate with a bot even when the bot was keen to cooperate. It’s not that we don’t trust the bot, it’s that we do: The bot is guaranteed benevolent, a capital-S sucker, so we exploit it.

That conclusion was borne out by conversations afterward with the study’s participants. “Not only did they tend to not reciprocate the cooperative intentions of the artificial agents,” Dr. Deroy said, “but when they basically betrayed the trust of the bot, they didn’t report guilt, whereas with humans they did.” She added, “You can just ignore the bot and there is no feeling that you have broken any mutual obligation.”

This could have real-world implications. When we think about A.I., we tend to think about the Alexas and Siris of our future world, with whom we might form some sort of faux-intimate relationship. But most of our interactions will be one-time, often wordless encounters. Imagine driving on the highway, and a car wants to merge in front of you. If you notice that the car is driverless, you’ll be far less likely to let it in. And if the A.I. doesn’t account for your bad behavior, an accident could ensue.

“What sustains cooperation in society at any scale is the establishment of certain norms,” Dr. Deroy said. “The social function of guilt is exactly to make people follow social norms that lead them to make compromises, to cooperate with others. And we have not evolved to have social or moral norms for non-sentient creatures and bots.”

That, of course, is half the premise of “Westworld.” (To my surprise Dr. Deroy had not heard of the HBO series.) But a landscape free of guilt could have consequences, she noted: “We are creatures of habit. So what guarantees that the behavior that gets repeated, and where you show less politeness, less moral obligation, less cooperativeness, will not color and contaminate the rest of your behavior when you interact with another human?”

There are similar consequences for A.I., too. “If people treat them badly, they’re programed to learn from what they experience,” she said. “An A.I. that was put on the road and programmed to be benevolent should start to be not that kind to humans, because otherwise it will be stuck in traffic forever.” (That’s the other half of the premise of “Westworld,” basically.)

There we have it: The true Turing test is road rage. When a self-driving car starts honking wildly from behind because you cut it off, you’ll know that humanity has reached the pinnacle of achievement. By then, hopefully, A.I therapy will be sophisticated enough to help driverless cars solve their anger-management issues.

Filed under Other Tagged with , , , , ,

Robots are coming and the fallout will largely harm marginalized communities

2021/05/01 Leave a comment

Interesting piece on the possible impact of automation on many of the essential service workers, largely women, visible minorities and immigrants (more speculative than data-driven):

COVID-19 has brought about numerous, devastating changes to people’s lives globally. With the number of cases rising across Canada and globally, we are also witnessing the development and use of robots to perform jobs in some workplaces that are deemed unsafe for humans. 

There are cases of robots being used to disinfect health-care facilities, deliver drugs to patients and perform temperature checks. In April 2020, doctors at a Boston hospital used Boston Dynamics’ quadruped robot called Spot to reduce health-care workers exposure to SARS-CoV-2, the virus that causes COVID-19. By equipping Spot with an iPad and a two-way radio, doctors and patients could communicate in real-time. 

In these instances, the use of robots is certainly justified because they can directly aid in lowering COVID-19 transmission rates and reducing the unnecessary exposure of health-care workers to the virus. But, as we know, robots are also performing these tasks outside of health-care settings, including at airports, offices, retail spaces and restaurants. 

This is precisely where the issue of robot use gets complicated. 

Robots in the workplace

The type of labour that these and other robots perform or, in some cases, replace, is labour that is generally considered low-paying, ranging from cleaners and fast food workers to security guards and factory employees. Not only do many of these workers in Canada earn minimum wage, the majority are racialized women and youth between the ages of 15 to 24. 

The use of robots also affects immigrant populations. The gap between immigrant workers earning minimum wage and Canadian-born workers has more than doubled. In 2008, 5.3 per cent of both immigrant and Canadian-born workers earned minimum wage, compared to 2018 where 12 per cent of immigrant workers earned minimum wage and only 9.8 per cent of Canadian-born workers earned minimum wage. Canada’s reliance on migrant workers as a source of cheap and disposable labour, has intensified the exploitation of workers.

McDonald’s has replaced cashiers with self-service kiosks. It has also begun testing robots to replace both cooks and servers. Walmart has begun using robots to clean store floors, while also increasing their usage in warehouses. 

Nowhere is the implementation of robots more apparent than Amazon’s use of them in its fulfilment centres. As information scholars applying marxist theory Nick Dyer-Witheford, Atle Mikkola Kjøsen and James Steinhoff explain, Amazon’s use of robots have reduced order times and increased warehouse space, allowing for 50 per cent more inventory in areas where robots are used, and have saved Amazon’s power costs by working in the dark and without air conditioning.

Already marginalized labourers are most affected by robots. In other words, human labour that can be mechanized, routinized or automated to some extent, is work that is deemed to be expendable because it is seen to be replaceable. It is work that is stripped of any humanity in the name of efficiency and cost-effectiveness for massive corporations. However, the influence of corporations on robot development goes beyond cost-saving measures. 

Robot violence

The emergence of Boston Dynamics’ Spot, gives us some insight into how robots have crossed from the battlefield into urban spaces. Boston Dynamics’ robot development program has long been funded by the American Defense Advanced Research Projects Agency (DARPA). 

In 2005, Boston Dynamics received funding from DARPA to develop one of its first quadruped robots known as BigDog, a robotic pack mule that was used to assist soldiers across rough terrain. In 2012, Boston Dynamics and DARPA revealed another quadruped robot known as AlphaDog, designed to primarily carry military gear for soldiers. 

The development of Spot would have been be impossible without these previous, DARPA-funded initiatives. While the founder of Boston Dynamics, Marc Raibert, has claimed that Spot will not be turned into a weapon, the company leased Spot to the Massachusetts State Police bomb squad in 2019 for a 90-day period. 

In February 2021, the New York Police Department used Spot to investigate the scene of a home invasion. And, in April 2021, Spot was deployed by the French Army in a series of military exercises to evaluate its usefulness on the future battlefield. https://www.youtube.com/embed/xYbhKHfZSEE?wmode=transparent&start=0Massachusetts State Police lease Boston Dynamics’ Spot in 2019.

Targeting the most vulnerable

These examples are not intended to altogether dismiss the importance of some robots. This is particularly the case in health care, where robots continue to help doctors improve patient outcomes. Instead, these examples should serve as a call for governments to intervene in order to prevent a proliferation of robot use across different spaces.

More importantly, this is a call to prevent the multiple forms of exploitation that already affect marginalized groups. Since technological innovation has a tendency to outpace legislationand regulatory controls, it is imperative that lawmakers step in before it is too late.

Source: https://theconversationcanada.cmail20.com/t/r-l-tltjihkd-kyldjlthkt-c/

Filed under Multiculturalism Tagged with , ,

← Older posts

Newer posts →