In another sign that the times, they are indeed a'changin', IBM released a letter from CEO Arvind Krishna to eight Democratic members of Congress on Monday that unequivocally pledges that Big Blue will no longer research, develop, nor sell facial recognition or analysis software. Said Krishna:
IBM firmly opposes and will not condone uses of any technology, including facial recognition freedoms, or any purpose which is not consistent with our values and Principles of Trust and Transparency. We believe now is the time to begin a national dialogue on whether and how facial recognition technology should be employed by technology offered by other vendors, for mass surveillance, racial profiling, violations of basic human rights and domestic law enforcement agencies.
The moral and ethical debate about facial recognition software has been heating up for the past two or three years. The AI-enabled technology itself has improved and been more widely deployed while the incidence of governments and individuals using it for nefarious reasons has grown. The Chinese use of the technology to closely monitor its Uighur-Muslim minority is a prime case in point.
Concerns about facial recognition software grew when research surfaced in 2018 that revealed the extent to which many commercial facial recognition systems (including those made by IBM) were seriously inaccurate in their identification of people of color. That finding was confirmed by a December 2019 National Institute of Standards and Technology study that found "empirical evidence for the existence of a wide range of accuracy across demographic differences in the majority of the current face recognition algorithms that were evaluated."
Facial recognition software is unreliable for law enforcement and security and a invitation for potential civil rights abuses. That view arises out of the fact that the software is provided by private companies with little regulation or federal oversight, and has confirmed bias problems in terms of age, race, and ethnicity, critics believe
Although the NIST study did not include Amazon's Rekognition program which is notable for its pursuit of law enforcement customers, the ACLU found that Rekognition incorrectly matched 28 members of Congress to faces picked from 25,000 public mug shots. Amazon has been notoriously aggressive in its pursuit of law enforcement contracts but it is only one og many players, most of which are not well known in the public domain.
Elsewhere, Facebook was ordered in January 2020 to pay $550 million to settle a class-action lawsuit over its unlawful use of facial recognition technology. And, of course, there is the Clearview AI caper, which we wrote about here.
IBM wants ‘national dialogue'
In his letter addressed to Sens. Cory Booker (D-NJ), Kamala Harris (D-CA), Reps. Karen Bass (D-CA), Hakeem Jeffries (D-NY), and Jerrold Nadler (D-NY) who have introduced a new House bill called The Justice in Policing Act of 2020, IBM's Krishna wrote:
IBM would like to work with Congress in pursuit of justice and racial equity, focused initially in three key policy areas: police reform, responsible use of technology, and broadening skills and educational opportunities…IBM welcomes your early leadership in announcing these proposals and stands ready to work with you and other Members of Congress, from both sides of the aisle, toward broad bipartisan legislation that can be enacted into law.
Krishna's letter spelled out three key suggestions the company believes would be helpful--new federal rules that hold police more accountable for misconduct, technology policies that increase transparency and help police protect communities but not promote discrimination or racial injustice, and training and education for in-demand skills to expand economic opportunity for communities of color. As examples, he suggested Congress consider scaling the P-TECH school model nationally and expanding eligibility for Pell Grants:
At IBM, we see an urgent demand for what we call "new collar" jobs, which require specialized skills but not necessarily a traditional 4-year college degree. Such jobs can still be found today in fast-growing fields from cybersecurity to cloud computing. We urge Congress to consider national policies to expand the number and reach of these programs.
IBM's sudden withdrawal from the market caught many industry insiders by surprise. As recently as last week, Cole Market Research issued a study with the rosy, and now outdated, title Image Recognition Market to Register Unwavering Growth During to 2025 - IBM, AWS
In November 2019, IBM called on Congress to regulate facial recognition with something it called "precision regulation." Since then local and state regulations have begun to bubble up everywhere and somewhere along the line, IBM has obviously now decided that being identified with the technology just isn't worth the tsuris.
As is the case with digital transformation, remote working, privacy, inequalities in income and access to health care, the combination of the pandemic lockdown and its new rules for working safely, the accompanying economic meltdown, the large and largely peaceful protests for equal justice, have accelerated the demand for change. Enterprises that have ignored or tried to bury difficult issues in the past suddenly find themselves at a place of reckoning. Lines are being drawn in the sand.
Skeptics will point out that IBM was already running a distant third to AWS and Microsoft in the facial recognition marketplace, that the company's announcement left loopholes that would allow it to sell the same software in the future in specific instances and even that it was a suspiciously timed exit designed to take advantage of the maximum PR value of being .the good guy at exactly the right time. Whatever its reasons, I'll take it