Will Big Tech be held accountable for its human rights abuses?
Welcome back from the weekend!
Corporations have the same rights as people now — even as they take rights away from the people they manipulate. …And we’re not even talking about the hundreds of thousands of murders that are directly attributed to Facebook’s deadly content policies. Or that Google is illegally vacuuming up personal patient data on hundreds of millions of Americans. Or that Amazon is officially lethal.
But regulatory and public sentiment is catching up…
This from Silicon Valley’s own backyard: “If Big Tech companies want to act like governments, they need to protect human rights, says the International Policy Director for Stanford University’s Cyber Policy Center
Cornell University’s J. Nathan Matias is admonishing researchers looking at any aspect of disinformation, population polarization or other aspect of problems related to Big Tech’s impact, to stop taking money from the companies being studied: “Why We Need Industry-Independent Research on Tech & Society”
At next month’s massive HIMSS healthcare technology gathering, The MITRE Corporation will tell healthcare providers and their technology suppliers to approach consumer-generated data cautiously — and offer recommendations for a model agreement for patient data collection and use
Two recent talks from the inaugural USF Center for Applied Data EthicsTech Policy Workshop position the situation a full-blown ethics crisis:
In “A filter and a spotlight: Technology and Mass Atrocity” (video), a former international human rights attorney and co-founder of Build Tech We Trust (created in the wake of the Gilroy, El Paso and Dayton shootings in the U.S) also talks about why the new group was created::
“We found ourselves really stunned by the rapid succession of terrorist attacks in the U.S. and the silence of our industry. We knew attacks were being coordinated on technology platforms and we kept waiting for leaders in tech to say something. And they didn’t. So we launched Build Tech We Trust”
In Building Empathy in Tech: A Critical Component for an Ethical Industry (video), the founder and executive director of the Tech Equity Collaborative (who also spent time at Code for America) details growing danger signs and offers suggestions for mitigation
And real consequences are starting to appear, led by investors, the courts, and healthcare regulators…
Google’s parent company is facing a major backlash from investors over its human rights policies. A dozen large shareholders — including Axa Investment Managers, the Church of England and Aviva Investors — have filed a resolution that accelerates the efforts of an even larger shareholder group that claims the company does not have proper controls over the way it monitors human rights risks. … Worth noting: This agitation could begin to explain the rash of recent C-suite firings at the company
A Dutch court ruled that a welfare surveillance system violates human rights
Germany’s health minister presented draft litigation for electronic patient record (EPR/EMR/EHR) data protection and use. This is in addition to the sweeping rules being crafted separately by the European Commission for overall technology strategy and governance
The U.S. Department of Health & Human Services announced that a new task force is looking at the Intersection of Clinical & Administrative Data (ICAD). The group will look at data standards for interoperability and prior authorization. Electronic medical records (EMR/EHRs) — created originally as a financial and administrative tool — are widely criticized for being deficient in clinical, and specifically patient-care, applications
TODAY IN…
DEEP TECH:
Turns out AI can’t actually distinguish human emotion
"‘A man’s face, which was bright red with his mouth open in a scream’ wasn’t expressing anger at all. It was a show of pure joy.”
Very important question: Are neuromorphic vision systems opening a Pandora’s Box?
On the side of the angels: Researchers are offering tips to help clinicians see when a machine learning study’s conclusions are suspect spanning methodology, data labels, and quality of data
Is the methodology specified in advance?In other words, did the researchers disclose ahead of time the outcomes they measured, along with sample size and the statistical methods used to analyze the data? If not, take any conclusions with a big grain of salt.
How were the data labels generated?If the labels used to characterize training data are based on expert opinions, as opposed to objective test results, then ground truth is much harder to establish. This can lead to low accuracy and cause the model to perform poorly in diverse clinical settings.
Was the model tested on external data?The machine learning system must be tested on data from an outside source to ensure its accuracy holds up in different geographies, populations, and clinical environments. That may seem obvious, but the researchers found that external validity was reported in only 6 of 18 deep learning studies they reviewed on the use of AI in image-based diagnosis.
MEDIA & TELECOM:
Twitter suspended 70 pro-Bloomberg accounts, citing platform manipulation
HEALTHCARE TECH:
The U.S. Department of Health & Human Services’ Office of the National Controller (ONC) issued two rulemaking documents in the run-up to next month’s HIMSS … See also: HIPAA enforcement appears to be headed in the wrong direction
Nine healthcare IT priorities for research will “align clinical and research technology infrastructure” and speed new medical research according to the ONC
Strategy on Reducing Regulatory and Administrative Burden Relating to the Use of Health IT and EHRs is billed as the “final” version of this ruling. The ONC has worked with the CMS since mid-2017 on the guidelines mandated by the 21st Century Cures Act
The second major health system this month is dropping Cerner and moving to Epic
FINANCE:
The private equity industry underperformed the S&P 500 over a 10-year period for the first time
Goldman Sachs pleaded not guilty to charges of misleading investors about the 1Malaysia Development Berhad (1MDB), in which former Goldman CEO Lloyd Blankfein has been directly implicated. Malaysia is seeing up to $7.5 billion in reparations.
The risk profile for the LBO market is… getting worse:
PUBLIC POLICY:
AI poses a test for the U.S. Patent Office over who should get credit as an inventor
The Justice Department is seeking outside legal help on its antitrust investigations against major U.S. tech companies
A watchdog group says the technology being used by the U.S. Census Bureau for the once-in-a-decade count is “not ready” … and the U.S. Government Accounting Office says the bureau faces “significant cybersecurity challenges”
And finally… The World Health Organization is lauding China’s draconian COVID19 measures, and Italy appears to be adopting China’s playbook. Now that the U.S. Centers for Disease Control has issued a blunt recommendation to hospitals to prepare for an outbreak, here’s a look at some of the '“best practices” China is establishing… and the spin cycle it’s directing at citizens:
China Ups Surveillance … See also:
China Turns to Tech Giants to Classify People During Coronavirus Outbreak … See also: