Algorithms, China, Congress, and us

Most of what you read in the news about Congress’s activities these days focuses on the debt limit, the infrastructure bill, reconciliation and generally passing bills to accomplish Biden’s legislative agenda to Build Back Better.

Oh, and then there is the surfacing of Facebook whistle blower Frances Haugen. This was the focus of a Senate hearing titled “Protecting Kids Online: Testimony from a Facebook Whistleblower” that took place on October 5. This was more interesting than I first expected. Just so you know, I am not on Facebook, never have been, never will be. [I do look at twitter almost daily, not a healthy habit.] What I thought was so interesting—and relevant to my job—is the shift that we are seeing in proposed legislation to curb problems with Facebook and other social media platforms. If I understand correctly, proposed legislation has typically attempted to regulate what could be posted.

We are now seeing bills introduced that, instead, would limit the further content the algorithms used are allowed to direct viewers to. One such piece of legislation is the Algorithmic Justice and Online Platform Transparency Act of 2021, introduced by Senator Markey (MA) and Congresswoman Matsui (CA, district 6). This bill would prohibit algorithms used by online platforms that discriminate on the basis of race, age, gender, ability and other protected characteristics. It would also take away the “black box” nature of current algorithms, by requiring online platforms to describe to users in plain language the types of algorithmic processes they employ and the information they collect. I know many in the math community work on this and many more care. Is it incumbent upon all of us to pay more attention? Obviously, there are others  in our community who have long been calling us to attention.

On the same date (October 5, lest you forget), the House Science Committee held a less sexy hearing titled “Balancing Open Science and Security in the U.S. Research Enterprise.” Competitiveness with China in particular is a hot topic in Washington, DC. This is not new.

Witness Maria Zuber of MIT is a co-chair of the recently constituted PCAST. She noted that National Security Presidential Memorandum (NSPM)-33—when released by the Trump administration—left unanswered questions and that clarifying information is being developed. She voiced concern “that increasingly federal agencies are attempting to impose limits on the kinds of research projects in which Chinese students can participate” and suggested that “limiting or discouraging Chinese students is the last step the U.S. should consider in countering China.” She urged that universities put in place their own clear and rigorous processes to review collaborations with China. At her home institution there is a process in place under which all collaborations with China, Russia and Saudi Arabia are reviewed.

Congressman Bill Foster of Illinois (PhD physicist) opened the hearing by observing that “Collaboration is the lifeblood of scientific discovery. Scientists build upon one another’s work, across time and around the globe. Openness in science allows reproduction and replication of work, increasing the reliability of conclusions and building public trust. It fosters cooperation across disciplines, brings in new perspectives, and sparks ideas that wouldn’t come from one solitary lab, or even one country.”

He noted that openness in the context of fundamental research does not come without risks and the goal of the hearing would be to explore what these risks are, the NSF’s workload associated with monitoring suspected cases of espionage, and with clarifying the procedures for researchers in the United States regarding propriety of foreign collaborations. Foster pointed out that these charges disproportionately target researchers of Chinese descent and that defendants with Chinese surnames were twice as likely to not be found guilty or to have all charges dropped, as compared to those with Western names.

During the hearing, Temple University’s Dr. Xiaoxing Xi’s personal story illustrated such a case and was especially moving. He recounted the day six years ago when armed FBI agents showed up at his door, raided his home, handcuffed him and led him away. Guns were pointed at his wife and young daughters. They were accusing him of passing sensitive technology to China. He told this story not to highlight his own innocence, but to point out that in such actions, the United States is not catching the real spies, and not using our tax money responsibly to protect our country. His testimony ends with

Let me be clear: a policy that targets Chinese scientists and cracks down on openness in fundamental research does not protect America’s research security. It makes the U.S. less competitive in innovation and less attractive to talents around the world. It threatens the U.S. leadership in science and technology. It must stop.”

One of my main jobs in Washington, DC (many would say my main job) is to advocate for increased funding for the NSF. One problem I see is that the agency is asked to do more and more, but funding is not provided for the staff to do so. Case in point is what we heard from another witness, NSF Inspector General Allison Lerner, who reported that investigating suspected cases of undue foreign influence is taking much of her office’s bandwidth.  Prior to late 2017, they had no foreign influence-related cases. As of October 4, 2021, such cases make up approximately 63% of her office’s caseload. They see only more cases coming at them, and have also experienced a dramatic increase in requests for assistance from the FBI since FY 2018, the vast majority of which relate to foreign influence issues. It now takes more than one full-time employee to process and assess these many requests.

The NSF also imposed a range of sanctions on the accused researchers—including suspending and terminating awards, and barring individuals from serving on panels and/or applying for grants in the future. Approximately 30 actions have been taken, and Lerner reported that the NSF has thus far recovered \$7.9 million from 23 grantees, at 21 institutions. According to an article in Science magazine, all but one of the cases involve an award to a scientist with links to China.

While we are seeing continuing scrutiny and concern, we are also seeing positive actions from the Biden administration. Congressman Foster made the tie to education at the hearing by saying that

“The United States attracts the largest share of international students worldwide, and three-quarters of them stay in the U.S. a decade after graduation, contributing to our economy. However, even before the COVID-19 pandemic, enrollment of international students was on the decline. And creating a hostile environment for foreign-born researchers will only exacerbate the problem. Asian-American advocacy groups, scientific societies, and universities have raised the alarm that prosecutions related to academic espionage have disproportionately targeted researchers of Chinese descent.”

Zuber’s statement notes that “According to the latest statistics from the National Science Foundation, 83 percent of Chinese students who received U.S. science and engineering doctorates between 2011 and 2013 were still in the U.S. five years later.”

I continue to be worried about all this anti-China in higher education stuff. When I read that dozens of Chinese students have been sent home from the University of Buffalo, or that Harvard is relocating a program from China to Taiwan (and these two are just from reading the news one morning!), I get a very chilly feeling. These are all pieces of a big puzzle.

The news isn’t all bad. I was pleased to see the recent statement of the Departments of State and Education issue a statement articulating “a renewed U.S. commitment to international education.” In this, math students are singled out—along with other science disciplines—for the contributions they make to benefit U.S. businesses upon graduation.

 

About Karen Saxe

Karen Saxe is Director of the AMS Office of Government Relations which works to connect the mathematics community with Washington decision-makers who affect mathematics research and education. Over many years she has contributed much time to the AMS, MAA, and AWM, including service as vice president of the MAA and in policy and advocacy work with all three. She was the 2013-2014 AMS Congressional Fellow, working for Senator Al Franken on education issues, with focus on higher education and STEM education. In Minnesota she has served on the Citizens Redistricting Commission following the 2010 census and serves on the Common Cause Minnesota Redistricting Leadership Circle. She has three children and, when not at work especially enjoys being with them and reading, hiking and sharing good food and wine and beer with family and friends.
This entry was posted in algorithmic bias, Congress, International science and tagged , . Bookmark the permalink.