This year’s Oscars was a production studded with women and people of color in presenter and award recipient roles. The event appeared to be deliberately crafted to display Hollywood’s acknowledgment that, yes, it has played an instrumental role in reinforcing white patriarchy. So, amid Hollywood’s exhibition of transformation, it seemed fitting when a multinational technology brand joined the virtue-signaling chorus.
Toward the end of the ceremony, IBM presented an open letter to “Tech,” in a commercial spot that almost immediately sparked criticism from AI ethics crusaders and tech industry skeptics.
In IBM’s “Dear Tech” ad, NASA hero Buzz Aldrin sat with a white board behind him, penning a missive to Tech, the main character in the commercial’s melodrama. “You’ve done a lot of good for the world,” Aldrin reminded Tech. Neuroscientist and TV nerd Miaym Bialik addressed Tech the way a concerned student advisor might: “I feel like you have the potential to do so much more,” she told it as pensive cello music sulked in the distance.
Critics pounced on IBM via Twitter. They alluded to examples of IBM practices that would not fit in Aldrin’s “lot of good for the world” category, some referencing an Intercept report that IBM employed NYPD camera footage data to develop facial recognition “features that allow other police departments to search camera footage for images of people by hair color, facial hair, and skin tone.”
IBM: "Can we build AI that fights bias?"
all of us: IBM you know we know about the whole surveillance thin-
IBM: Nah, we're actually asking. Does anybody know? #ai #ethics #oscars— Blakeley H. Payne (@blakeleyhpayne) February 25, 2019
IBM’s ad seemed to be saying Tech has gotten some bad report cards lately. It’s time to get back on track — hence, this commercial intervention. Others in the ad with less familiar faces emerged, people with ideas for how Tech can improve. Open Source software movement co-founder Bruce Perens told Tech, “Let’s champion data rights as human rights.” MeToo movement founder Tarana Burke asked, “Are you working for all of us or just a few of us?”
MIT researcher and facial recognition activist Joy Buolawmini wasn’t buying IBM’s reflective stance. “We must remain watchful of all tech companies and hold them accountable when they negatively impact society,” she told RedTail.
She and others who resented what they saw as IBM’s ethical façade took action. Within a week, Buolawmini had produced, written and directed a response video parodying the IBM ad. The “Dear Tech Company” video features a collective of academics and advocates for democratic tech, people whose books and other work Buolawmini said have “deepened my understanding of the limitations of technology and how tech companies can be complicit in exacerbating inequalities if left unchecked.”
IBM did not respond to a request to comment for this story.
“Ultimately, ethical statements must be backed by consistent actions.”
– Joy Buolawmini.
In the Dear Tech Company parody, Director of Civic Media at MIT Ethan Zuckerman sits in front of a white board like Aldrin did in the original ad. Rather than mathematical equations, his white board lists names of books exposing algorithm-enabled problems: Cathy O’Neil’s Weapons of Math Destruction, Safiya Noble’s Algorithms of Oppression and Ruha Benjamin’s Race After Technology: Abolitionist Tools for the New Jim Code. Noble and Benjamin are also featured in the Dear Tech Company video.
Buolawmini has taken the AI industry to task many times already, particularly in the arena of facial recognition algorithms. In research published about a year ago, she found that IBM’s facial recognition system had an error rate of nearly 35 percent when attempting to detect darker-skinned women. A few months later, IBM released its “The Diversity in Faces” dataset, its answer to improving demographic diversity in data used to train facial recognition systems.
For the record, Buolawmini said IBM did not ask her to appear in its Dear Tech ad.
The Brand Metrics of Cause Marketing
The advertising industry considers the Oscars “The Super Bowl for women,” an appointment-television event garnering big ratings and splashy branding opportunities that will reach one of the year’s largest audiences. In an article about this year’s Oscars sponsors, IBM’s VP of Corporate Marketing Ann Rubin told Variety that the company ran an Oscars ad in 2017, noting, “according to our brand metrics, it actually worked very well.”
During a TV event in which the influential film industry embraced the people it has dismissed for so long, here was an equally influential tech brand joining along with a similar strategically-designed marketing message intended to engender trust. Like a remorseful teen slacking off during senior year, IBM seemed to say, “We will do better. We promise. Trust us!”
But feel-good corporate branding sometimes belies business behavior with different priorities. Despite IBM’s apparent support for Perens’s open data message, the company and the tech industry as a whole has enabled the commodification of our data, turning it into massive revenue streams. IBM’s Cloud services, which provide a virtual home for all that digitized data, brings in billions of dollars and is a growing portion of the company’s business.
Will IBM’s definition of transparency be the same one that democratic data rights advocates use?
Indeed, when IBM claims in its “Principles for Trust and Transparency” that “Data and insights belong to their creator,” the company isn’t referring to individual people as owners of their data at all. Instead, the company promises, “IBM clients’ data is their data, and their insights are their insights.” In this framework, data rights are for corporations. IBM wants to ensure its corporate clients that they needn’t worry that IBM will use their data or relinquish it to prying parties like governments.
Government regulators, however, are among the very entities that one day may request information about algorithms and the data used to train them. Will IBM’s definition of transparency be the same one that democratic data rights advocates use?
Soon after the Oscars, RIT Philosophy Professor Evan Selinger, also a collaborator in the parody video, criticized the IBM ad in a Slate opinion piece, suggesting that when tech firms say they support certain principles, they don’t always have the same thing in mind as activists might. “A commercial like this one can’t avoid being an empty marketing pitch when it represents a contested concept as a clear and unambiguous wish that technology can magically grant just as easily as Santa can satisfy a request for a new smartphone,” he wrote.
The IBM ad is likely the first on TV to mention the concept of AI bias, indicating the issue has reached the mainstream consciousness.
It’s worth noting the IBM ad is likely the first on TV to mention the concept of AI bias, indicating the issue has reached the mainstream consciousness. In the ad, people including musician and tech critic Janelle Monae asked, “Can we build AI without bias? AI that fights bias? AI that helps us see the bias in ourselves?”
In its trust and transparency principles, the company promises to ensure that data used to train AI aligns with human values and expectations, to make AI transparent and allow for explainable AI decisions. Yet in doing so, IBM stresses that its clients’ intellectual property is core to this mission. The policy emphasizes “The principle that clients own their own business models and intellectual property and that they can use AI and cognitive systems to enhance the advantages they have built.” Again, whether IBM’s definition of transparency and explainable AI is the same as that of advocates for data rights and access is questionable.
Expect More Displays of Virtue from Big Tech
Corporate tech senses the winds of change blowing. Employees of Microsoft and Google are rising up against military use of the systems they’ve helped develop. Labor groups fought Amazon in New York for its poor record on workers’ rights. Meanwhile, media reports remind us every day that Facebook and the data industrial complex are compromising our privacy and profiting from surveillance capitalism.
Corporations like IBM have massive advertising and marketing budgets and troupes of PR professionals behind them. Their inclination is to get out in front of issues that could soil brand reputation. As tech’s once-inspiring story of innovation and new tomorrows becomes besmirched by news of unethical practices and fear of AI gone awry, we can expect to see more companies attempt to steer the conversation in a way that presents their technologies in a positive light.
As tech’s once-inspiring story of innovation and new tomorrows becomes besmirched by news of unethical practices and fear of AI gone awry, we can expect to see more companies attempt to steer the conversation in a way that presents their technologies in a positive light.
But this sort of marketing approach runs the risk of being construed as hypocritical gimmickry not unlike the so-called green-washing perfected by oil industry advertising: touting clean energy investments while lobbying behind-the-scenes in support of fossil fuels.
“Ultimately, ethical statements must be backed by consistent actions,” said Buolawmini. She alluded to the final scene of the Dear Tech Company video, in which she lifts her arms to draw attention to the letters “ICU” hovering above her in a striped font much like the one used in the IBM logo. “The final message of the video is that as tech companies increasingly watch us through more data collection we are watching back. We see you. I see you. I-C-U.”
0 comments on “To Ethical Tech Advocates, IBM’s “Dear Tech” Ad Looks Like Moral Posturing”