Making AI Less Squirrelly

You may have missed it, but the Association for the Advancement of Artificial Intelligence (AAAI) just announced its first annual Squirrel AI award winner: Regina Barzilay, a professor at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL).   In fact, if you’re like me, you may have missed that there was a Squirrel AI award.  But there is, and it’s kind of a big deal, especially for healthcare – as Professor Barzilay’s work illustrates. 

The Squirrel AI Award for Artificial Intelligence for the Benefit of Humanity (Squirrel AI is a Chinese-based AI-powered “adaptive education provider”) “recognizes positive impacts of artificial intelligence to protect, enhance, and improve human life in meaningful ways with long-lived effects.”  The award carries a prize of $1,000,000, which is about the same as a Nobel Prize. 

Yolanda Gil, a past president of AAAI, explained the rationale for the new award: “What we wanted to do with the award is to put out to the public that if we treat AI with fear, then we may not pursue the benefits that AI is having for people.”

Dr. Barzilay has impressive credentials, including a MacArthur Fellowship.   Her expertise is in natural language processing (NLP) and machine learning, and she focused her interests on healthcare following a breast cancer diagnosis.  “It was the end of 2014, January 2015, I just came back with a totally new vision about the goals of my research and technology development,” she told The Wall Street Journal. “And from there, I was trying to do something tangible, to change the diagnostics and treatment of breast cancer.”

Since then, Dr. Barzilay has been busy.  She’s helped apply machine learning in drug development, and has worked with Massachusetts General Hospital to use A.I. to identify breast cancer at very early stages.  Their new model identifies risk better than the widely used Tyrer-Cuzick risk evaluation model, especially for African-American women. 

As she told Will Douglas Heaven in an interview for MIT Technology Review:  “It’s not some kind of miracle—cancer doesn’t grow from yesterday to today. It’s a pretty long process. There are signs in the tissue, but the human eye has limited ability to detect what may be very small patterns.”

This raises one of the big problems with AI; we may not always understand why AI made the decisions it did.  Dr. Barzilay observed:

But if you ask a machine, as we increasingly are, to do things that a human can’t, what exactly is the machine going to show you? It’s like a dog, which can smell much better than us, explaining how it can smell something. We just don’t have that capacity.

She firmly believes, though, that we can’t wait for “the perfect AI,” one we fully understand and that will always be right; we just have to figure out “how to use its strengths and avoid its weaknesses.”   As she told Stat News, we have a long way to go: “We have a humongous body of work in AI in health, and very little of it is actually translated into clinics and benefits patients.”

Dr. Barzilay pointed out: “Right now AI is flourishing in places where the cost of failure is very low…But that’s not going to work for a doctor… We need to give doctors reasons to trust AI. The FDA is looking at this problem, but I think it’s very far from solved in the US, or anywhere else in the world.” 

A concern is what happens when A.I. is wrong.  It might predict the wrong thing, fail to identify the right thing, or ignore issues it should have noticed.  In other words, the kinds of things that happen every day in healthcare already.  With people, we can fire them, sue them, even take away their license.  With A.I., what we do to whom/what is not at all obvious.

“This is a big mess,” Patrick Lin, director of Ethics and Emerging Sciences Group at California Polytechnic State University, told Quartz. “It’s not clear who would be responsible because the details of why an error or accident happens matters.” 

Wendall Wallace, of Yale University’s Interdisciplinary Center for Bioethics, added: “If the system fails to perform as designed or does something idiosyncratic, that probably goes back to the corporation that marketed the device.  If it hasn’t failed, if it’s being misused in the hospital context, liability would fall on who authorized that usage.”

“If it’s unclear who’s responsible, that creates a gap, it could be no one is responsible,” Dr. Lin said. “If that’s the case, there’s no incentive to fix the problem.”  Oh, great, just what healthcare needs: more unaccountable entities.

To really make AI succeed in healthcare, we’re going to have to make radical changes in how we view data, and in how we approach mistakes.

AI needs as much of data as it can get.  It needs it from diverse sources and on diverse populations.  All of those are problematic in our siloed, proprietary, one-step-from-handwritten data systems.  Dr. Barzilay nailed it: “I couldn’t imagine any other field where people voluntarily throw away the data that’s available. But that’s what was going on in medicine.” 

Despite our vaunted scientific approach to medicine, the fact is that we don’t really know what happens to most people most of the time, and do a poor job of counting even basic healthcare system interactions, like numbers of procedures, adverse outcomes, even how much things cost.  As bad as we are at tracking episodic care, we’re even worse at tracking care -- much less health -- over time and across different healthcare encounters. 

Once AI has data, it is going to start identifying patterns, some of which we know, some of which we should have known, and some of which we wouldn’t have ever guessed.  We’re going to find that we’ve been doing some things wrong, and that we could do many things better.  That’s going to cause some second-guessing and finger-pointing, both of which are unproductive.

Our healthcare system tends to have its head in the sand about identifying errors/mistakes, for fears about malpractice suits (justified or not).  Whatever tracking does happen is rarely disclosed to the public.  That’s a 20^th century attitude that needed to be updated in an AI age; we should be thinking less about a malpractice model and more about a continuous quality improvement model.

“The first thing that's important to realise is that AI isn't magic,” David Champeaux of Cherish Health said recently.  It’s not, but neither is what we already do in healthcare.  We need to figure out how to demystify them. 

WeChat to Many, But WeDoctor to Some

You’ve probably heard about TikTok, especially lately.  President Trump wanted a ban on it, and seems to have endorsed a deal for a U.S.-based version of it.  The hundred million U.S. users, and probably their parents, are undoubtedly watching the sequence of events with mixed amusement and concern. 

But you may have paid less attention to what’s been going on with WeChat, another China-based app.  WeChat was part of the original proposed ban, which a federal judge blocked this weekend, hours before it was due to go into effect (the Commerce Department plans to appeal).  The ban is on “transactions,” which, in WeChat’s case, covers a lot of ground. 

TikTok was overlooked by authority figures for a long time because it was mostly used by young people and mostly for what seemed, to them, to be trivial purposes.  Not so with WeChat; it is deeply engrained in users’ lives, including for their health.   

WeChat is owned by Tencent Holdings, one of China’s internet giants.  It has been described as a “Swiss Army knife” app, able to do many tasks – not just messaging and social networking, but also games, shopping, and payments.  You can order food or book travel.  For many users it is a primary source of news, which is part of the problem. 

It is also important to users’ health.  WeChat is, according to CMI Media, “fast becoming the #1 online healthcare destination in China.”  It offers, among other things, health content (some in partnership with U.S. firms), health products, telehealth, a network of “trusted” doctors, a form of health insurance, and WeDoctor.  The latter “provides online health enquiry service, psychological support, prevention guidelines and real-time pandemic reports,” and is free to the user.  It is available “24/7 for people all over the world.”

Most notably, WeDoctor is preparing for an IPO for late this year/early 2021, which could value it as high as $10b.  I would again note the “24/7 for people all over the world.”

If we’re worried about what information China might glean from the video-watching habits of teenagers, think about how worried we should be about China having access to what health information users sought, what medical advice they got, and what health products they ordered.

China is famed for its “Great Firewall,” which restricts which outside internet platforms – like Google or Facebook – can be used within its borders.  Equally important, the Chinese government monitors what happens on WeChat and other internet platforms/apps, and does not allow news or opinions it finds objectionable, or subversive.  You might think you are in your own Facebook or Twitter bubble in the U.S., but in China – or on WeChat – that bubble is shaped and controlled by the government. 

As a result, Politico reported, “Now young online Chinese, once conduits for new ideas that challenge the power structure, are increasingly part of Beijing’s defense operation.”  Even U.S. users find their worldview shaped by the content they are allowed to see.  As The New York Times said, “it has helped bring Chinese censorship to the world.”

“All of a sudden I discovered talking to others about the issue didn’t make sense,” one user told The New York Times.  “It felt like if I only watched Chinese media, all of my thoughts would be different.”

There are estimated to be 19 million U.S. users, out of WeChat’s 1.2 billion users; most are people with family or friends in China, who rely on the app to stay in touch.  The U.S. may argue it is worried about what financial and personal information might be going to the Chinese government, but it should be equally worried about what “information” is being served to U.S. users. 

Think, for example, what it might tell U.S. users about COVID-19 vaccines.

The U.S. moves make some worry that we’re becoming more like China, leading to the “splinternet” where, as Vox explained, “your experience of the internet increasingly depends on where you live and the whims of the ruling parties there.” 

Vox goes on to note:

Nations are increasingly pursuing various forms of internet sovereignty, from Russia building a walled-off intranet to India regularly shutting down the internet in areas of social unrest to some European nations introducing a right to be forgotten from search engines.

It is the opposite of the open access, no borders version of the internet that most of us have believed in for the past thirty years.  Aaron Levie, CEO of cloud-computing company Box Inc, warned in The Wall Street Journal: “U.S. tech companies have far more to lose if this becomes a precedent.  This creates a Balkanization of the internet and the risk of breaking the power of the internet as one platform.”

One Congressional official told Wired:

We are finally having the debate China had two decades ago, when it put in the Great Firewall because it found foreign technology threatening its political system. Only now is America catching up with foreign technology that is a direct threat to our open system.

But Jason Healey, an expert on cyber conflict, competition and cooperation at Columbia University, told The New York Times: “The vision for a single, interconnected network around the globe is long gone. All we can do now is try to steer toward optimal fragmentation.”

Somehow, “optimal fragmentation” isn’t how I want to think of my internet experience; I suspect that fragmentation won’t be so optimal.

Even if some version of the ban on WeChat goes through, it’s not clear how effective it would be.  Options like using VPNs or downloading the app from non-store channels may allow users to continue to use it.  In any cases, The Washington Post reported that “the administration does not intend to prosecute anyone for finding new ways to use the apps.” 

In discussing the effect of potential WeChat bans with The New York Times, Fang Kecheng, a professor at the Chinese University of Hong Kong, said: “Information is like water. Water quality can be improved, but without any flow, water easily grows fetid.”  He didn’t carry the analogy further, but I will: information is like water, in that, eventually, it will get to where it wants to go. 

We don’t have a U.S. platform as versatile as WeChat; we don’t even have a health platform as capable as WeChat’s health capabilities.  But, if we’re not careful, WeChat might become that platform.   

Healthcare Can Learn from Chess

 Oh, gosh, two of my favorite things are in the news together: Twitch and chess. 

Just kidding.  I barely know what Twitch is, and the last time I played chess was, well, not in this century (and, even then, not well).  But I’m not kidding about their convergence.  Chess has become a big hit on Twitch, especially in these COVID times. 

I figure, if two such seemingly divergent things are meshing, there must be some lessons there, even for healthcare. 

For those of you over, say, fifty, Twitch is an online service that facilitates livestreaming, particularly of gaming.  That is, people watch other people playing games, such Fortnite or League of Legends. 

E-sports, as this is known, have become a big thing; colleges are even giving out scholarships for e-sports.   Major news outlets, such as The New York Times and The Washington Post, reported on Twitch re-signing video game star Tyler Blevins, a.k.a “Ninja,” much as they might have reported an NFL team signing a star player. 

As I write, 2.7 million people are livingstreaming on Twitch.  Its all-time concurrent viewers peak is just over 6 million.  There were 1.6 billion hours watched in August, with over 11 billion year-to-date.  It draws more viewers than network television hits. There are 93,000 live channels at this moment. 

Some of those livestreams are watching chess.  Analytics website Sullygnome reports people watched 45 million hours of chess on Twitter in 2020.  In June an amateur chess tournament was (briefly) the top-viewed Twitch stream. The convergence of streaming and chess has created a “giant chess bonfire,” Marcus Graham, Twitch’s head of creator development, told The New York Times. 

The biggest Twitch chess star is grandmaster Hikaru Nakamura, who accounted for 10 million of those 45 million hours watched.  Last month esports organization Team SoloMid signed Mr. Nakamura to what is believed to be a six figure deal; he is not the only chess player signing such deals, nor is it the only such organization signing them.  Other key players include Alexandra Botez, Eric Hansen, and Felix “xQc” Lengyel (who is more famous for other games). 

Mr. Nakamura has over 500,000 Twitch followers, drawn as much by how he engages viewers as his chess expertise.  “He draws people because he’s so good, but also, there are other top players on Twitch that are not as engaging as he is, not as funny, not as in tune with the sort of Twitch culture,” one viewer told The Times.

Mr. Nakamura’s appeal is, in part, because he helps demystify the game; “Fans love when he loses his mind at a poorly thought-out move.”  One fan told The Times:

When I was growing up, high-level chess was secreted behind closed doors, played by the privileged, moneyed people in society.  Realizing that I could see chess being played at the highest levels and seeing the players on Twitch communicate their thoughts with the community really resonated with me.

In an interview with NPR. Mr. Nakamura acknowledged the conflict: “So there are people who feel very much that chess, as it is, should stay that way as opposed to, say, in a sense cheapening the game by trying to make it more popular.”  His goal is to help break down those barriers.

As The Times asked, “If Serena Williams and Usain Bolt showed off their unique abilities every day on a livestream, wouldn’t you watch?” 

Lest we think this is entirely about Twitch, I’ll point out that Chess.com is the world’s third largest gaming website, with 42 million members and an average of over 9 million regular viewers.  It claims to be adding over 1 million users per month.  Chess is adapting to the times.

Twitch isn’t just focused on games, much less on chess.  It has signed musicians and other creators.  It livestreams sports and concerts.  It uses its connection to Amazon Prime to host Watch Parties, so users can “watch, react, and discuss any movie or TV show that is available with their Amazon Prime or Prime Video subscription, directly on Twitch.”  YouTube and Netflix should be worried. 

I actually wrote about Twitch several years ago, just after Amazon acquired it for close to a billion dollars, noting at the time:

I'll posit this: if you didn't know what Twitch was, or are still puzzled at its appeal, your organization is going to have a tough time competing over the next couple of decades.

That’s even more true today than ever.

As with TikTok, Twitch users are predominately young.  Forty-one percent of its users are 16-24, with another 32% 25-34.  Only 3% are 55-64; users over 65 don’t even count as a rounding error.  It is the converse of our healthcare system’s current users, but it is the generational wave that is coming. 

I’ve written before about how healthcare has lessons to learn from disparate platforms like TikTok or Fortnite, as those platforms have extended their reach (e.g., social justice and the Metaverse, respectively).  As best I can tell, Twitch hasn’t paid any attention to health or healthcare.  Then, again, gaming/gamification are still struggling to gain a foothold in healthcare. 

Healthcare likes its traditions – white coats, anyone? – and claims to value our privacy so much that it often makes it hard for us or our caregivers to get access to our information.  “Games” or livestreaming to a participatory audience seems like the remotest thing for healthcare.

But if the staid, traditional-laden of chess can thrive in a Twitch environment, there have to be ways that healthcare can take advantage of it as well.  Too much of what goes in healthcare is “secreted behind closed doors,” as the one chess fan said about chess.  We need to open those doors.

I’m not talking about easy uses like yoga classes or even medical lectures; I’m talking about more things like livestreaming surgeries, as Tokyo Women’s Medical University does – in VR.   COVID-19 is prompting some medical schools to do the same.  There are uses we have not yet thought of, but should.  It just probably won’t be healthcare’s largely Baby Boomer leadership that will come up with them. 

What healthcare organization or professional will have the first breakout Twitch channel?  

A New Kind of Labor Day

This is probably the strangest Labor Day in decades, perhaps ever.   Tens of millions of workers remain unemployed due to the COVID-19 pandemic.  Many of those who are still working are adapting to working from home.  Those who are back at their workplace, or never left, are coping with an array of new safety protocols. 

Those who work in the right industries – like the NBA – may get tested regularly but most workers have to figure out for themselves when to quarantine and when to get tested.  For many workers, such as health care workers, people of color, and workers with underlying health issues, going to work is literally a life-or-death calculation. 

No wonder that experts, like Dr. David B. Agus, are calling for companies to have Chief Health Officers. 

Labor Day was originally intended to celebrate the labor movement, but these days labor unions don’t have much to celebrate.  Only around 10% of U.S. workers belong to a labor union; both the number and the percent of unionized workers has been in steady decline over the past few decades. 

Now Labor Day is mainly an extra day off for most, the unofficial end to summer, and, this year, possibly the springboard to a new surge in COVID-19 cases, due to holiday celebrations.  Dr. Anthony Fauci warned:

We don’t want to see a repeat of the surges that we have seen following other holiday weekends.   We don’t want to see a surge under any circumstances, but particularly as we go on the other side of Labor Day and enter into the fall.

Eleanor Murray, an assistant professor of epidemiology at the Boston University School of Public Health, explained to The New York Times:

People are getting tired of taking these precautions and of having their lives upended.  They’re missing their friends and family, and everyone wishes things were back to normal. That’s totally understandable, but unfortunately we don’t get a say, really.

Las Vegas Labor Day party

It doesn’t help that many schools have reopened, with colleges and universities already seeing mass outbreaks.  One estimate found at least 25,000 such cases, in 37 states.  Labor Day parties and gatherings could send that number soaring.  No one is going to be surprised by photos of swarms of beachgoers, pool parties, and crowded bars, and no one can be surprised if this fall sees the consequences of that in more COVID-19 cases, hospitalizations, and deaths.

Many wonder if the nature of work will change as a result of the pandemic.  Most companies have been pleasantly surprised to find that productivity hasn’t suffered by work from home.  A survey by Mercer found that 94% of employers said productivity was the same (67%) or even higher – and 73% said they expected at least a quarter of their workforce would remain working remotely. 

Workers are generally happy too.  For one thing, they’re saving the equivalent of “billions” of dollars in avoided commuting time, not to mention actual costs such as gasoline, parking, and mileage.  Some see “digital nomads” as the future.  A June PwC survey found that, even post-pandemic, 83% of office workers wanted to work at least one day a week from home, and 55% of employers thought most of their workers might do so. 

Not everyone is a fan of remote work.  Reed Hastings, founder and co-CEO of Netflix, called it a “pure negative,” citing the benefits of getting together in person.  Still, even he concedes the new balance might be four days in the office, one day virtual. 

Mr. Hastings is not alone in his concerns about the benefits of in-person interactions, the so-called “water cooler” or “hallway” effect – serendipitous encounters that have been linked to creative thinking.  “These ties are critical to our well-being because they end up giving us the opportunity to vent, confide, brainstorm, and discuss things that we think are important,” says Mario Luis Small, a Harvard sociology professor. 

"The water cooler as a place to build relationships just evaporated when everybody went remote,” Dan Manian, CEO of Donut, told The Wall Street Journal.  Professor Eddie Obeng also told The Wall Street Journal, “employees are using tools and methods that weren’t designed for the current world in which we are living.”  Donut is one of the companies trying to provide solutions for this problem. 

Slack, Microsoft Teams, and Zoom are also trying to do the same, as are less well known companies like Hallway, Miro, Qube, Sidekick, and Spatial.  .

We need to be thinking bigger.  As Dr. Agus said in a May interview with Salesforce:

I want companies to have a program where they think about their employees’ productivity and health every day. And, then, if they're customer-facing, how do they convey that same message to the individual customer? We have to think differently in that way.

Some companies and other organizations are starting to take Dr. Agus’s suggestion seriously, such as Bowling Green State University.  Cambridge Health advisors, which aims at retailers and restaurants, urges: “New C-level leadership, a chief health officer, is needed to establish a vision for chainwide public health and to ensure store level initiatives are executed timely and effectively.” 

Adam Aron, CEO of AMC Theaters, says frankly:

The single biggest issue facing businesses in the United States is how do we manage our way through the coronavirus crisis.  The CEO of every major company in the country is going to have to make public health the single top vision of the company.”

Imagine that.

It shouldn’t have taken a pandemic to greatly increase remote work.  It shouldn’t have taken a pandemic for companies to think about their employees’ productivity and health every day (and to weigh those at least equally).  It shouldn’t have taken a pandemic to make public health top-of-mind for CEOs of every major company.  But it has, and, even now, it’s not clear how many companies will actually change long term.

It’s been about six months now since the restrictions/quarantines/lockdowns started, with the resulting job losses and economic havoc.  We’ve learned a lot, at a terrible cost, but it’s anyone’s guess about what the next six months will bring (especially as flu season is about to start).

So, let’s celebrate Labor Day – responsibly! – and hope that we’ve learned the right lessons.