1 big thing: The return of analog?

2. The privacy smokescreen

Illustration of ripped out paper that reveals 0’s and 1’s.

3. Automation bias on the move

A woman in a car with a phone in one hand adjusts the radio with her other.

4. Worthy of your time

Illustration of a nearly-empty bottle of cleaning solution with the Facebook logo on it.

5. 1 fun thing: How SF stepped on a rake

A man uses facial recognition on an iPhone

Welcome back to Future! This is our last letter of 2019 — and my final issue as your correspondent.

It's been such a delight bringing you stories from the leading edge of technology and hearing back from you in return.

I'm off to something new in January. Erica Pandey (erica@axios.com) will continue to bring you smart writing and deep reporting in the new year.

This issue is 1,300 words, or a 5-minute read.

Future

A robotic hand with the palm facing upward.

deepBlue

Axios Future

Standard

1 big thing: AI washing

Illustration of a woman working at a computer wearing a robot mask

2. An uncertain future for Big Beer

Illustration of a mug of beer, with a sad face in the foam

3. Tech giants reach further into health care

Apple's Stan Ng speaks on a stage in front of photos of an Apple Watch

Illustration of $100 bills being twirled on a fork. 

5. 1 not-fun thing: Dopamine fasting

A man sitting alone in a chair looking out the window

Welcome to Future! Thanks for reading. Get in touch by replying to this message or at kaveh@axios.com. Erica, who writes Future on Wednesdays, is at erica@axios.com.

This issue is 1,273 words, a 5-minute read. 

1 big thing: U.S. takes aim at Chinese drones…

Illustration of the Chinese flag with drones in place of stars

2. … leaving a potential vacuum

A DJI drone hovering in front of a background with the DJI logo

3. What the future of work looks like today

A sign in a Walmart pointing customers toward a self checkout

Illustration of a person's outline filled in with 1s and 0s, on a yellow background

5. 1 fun thing: Uncovering an ancient glyph with AI

A glyph of an animal shape with large hands

This issue is 1,477 words, < a 6-minute read.

1 big thing: An obstacle course for AI

Illustration of a robot jumping over a hurdle

2. The new AI co-writer

Illustration of a robot hand holding a quill.

3. Cars have become computers 

Illustration of a car made out of circuitry 

Illustration of a dollar bill disintegrating into pixels

5. 1 fun thing: NYC's linguistic landscape

A photo of museum brochures in various languages

Welcome back to Future. Thanks for reading! Get in touch by replying to this email or writing me at kaveh@axios.com. Erica, who writes this newsletter on Wednesdays, is at erica@axios.com.

📺 In the first “Axios on HBO” special, Joe Biden accuses the media of misjudging how liberal the Democratic Party really is and dismisses the idea that Rep. Alexandria Ocasio-Cortez defines it.

Catch a sneak peek and watch the full interview this Sunday at 6:30 p.m. ET.

This issue is 1,389 words, ~ a 5-minute read.

1 big thing: A tug-of-war over biased AI…

Illustration of computer with a frowning face.

2. …and how the split shapes AI's uses

Illustration of computer with a smiley face.

3. Counter-drone tech explodes

A man crouches in military fatigues and points a rifle-like device toward the sky

Illustration of a robot holding a phone whose cord is being cut by a human holding scissors

5. 1 fun thing: Science tackles an uncanny mystery

Welcome back to Future! Thanks for reading. Please get in touch: You can reply to this email or send me a note at kaveh@axios.com. My colleague Erica Pandey, your Wednesday Futurist, is at erica@axios.com.

This issue is 1,342 words, a 5-minute read.

National

The idea that AI can replicate or amplify human prejudice, once argued mostly at the field's fringes, has been thoroughly absorbed into its mainstream: Every major tech company now makes the necessary noise about "AI ethics."

Yes, but: A critical split divides AI reformers. On one side are the bias-fixers, who believe the systems can be purged of prejudice with a bit more math. (Big Tech is largely in this camp.) On the other side are the bias-blockers, who argue that AI has no place at all in some high-stakes decisions.

Why it matters: This debate will define the future of the controversial AI systems that help determine people's fates through hiring, underwriting, policing and bail-setting.

What's happening: Despite the rise of the bias-blockers in 2019, the bias-fixers remain the orthodoxy.

A recent New York Times op-ed laid out the prevailing argument in its headline "Biased algorithms are easier to fix than biased people."

"Discrimination by algorithm can be more readily discovered and more easily fixed," says UChicago professor Sendhil Mullainathan in the piece. Yann LeCun, Facebook's head of AI, tweeted approvingly: "Bias in data can be fixed."

But the op-ed was met with plenty of resistance.

The other side: At the top academic conference for AI this week, Abeba Birhane of University College Dublin presented the opposing view.

Birhane's key point: "This tool that I'm developing, is it even necessary in the first place?"

She gave classic examples of potentially dangerous algorithms, like one that claimed to determine a person's sexuality from a photo of their face, and another that tried to guess a person's ethnicity.

"[Bias] is not a problem we can solve with maths because the very idea of bias really needs much broader thinking," Birhane tells Axios.

The big picture: In a recent essay, Frank Pasquale, a UMD law professor who studies AI, calls this a new wave of algorithmic accountability that looks beyond technical fixes toward fundamental questions about economic and social inequality.

"There's definitely still resistance around it," says Rachel Thomas, a University of San Francisco professor. "A lot of people are getting the message about bias but are not yet thinking about justice."

"This is uncomfortable for people who come up through computer science in academia, who spend most of their lives in the abstract world," says Emily M. Bender, a University of Washington professor. Bender argued in an essay last week that some technical research just shouldn't be done.

The bottom line: Technology can help root out some biases in AI systems. But this rising movement is pushing experts to look past the math to consider how their inventions will be used beyond the lab.

"AI researchers need to start from the beginning of the study to look at where algorithms are being applied on the ground," says Kate Crawford, co-founder of NYU's AI Now Institute.

"Rather than thinking about them as abstract technical problems, we have to see them as deep social interventions."

The impact: Despite a flood of money and politics propelling AI forward, some researchers, companies and voters hit pause this year.

Most visibly, campaigns to ban facial recognition technology succeeded in San Francisco, Oakland and Somerville, Mass. This week, nearby Brookline banned it, too.

One potential outcome: freezes or restrictions on other controversial uses of AI. This scenario scares tech companies, who prefer to send plumbers in to repair buggy systems rather than to rip out the pipes entirely.

But the question at the core of the debate is whether a fairness fix even exists.

The swelling backlash says it doesn't — especially when companies and researchers ask machines to do the impossible, like guess someone's emotions by analyzing facial expressions, or predict future crime based on skewed data.

"It's anti-scientific to imagine that an algorithm can solve a problem that humans can't," says Cathy O'Neil, an auditor of AI systems.

These applications are "AI snake oil," argues Princeton professor Arvind Narayanan in a presentation that went viral on nerd Twitter recently.

The main offenders are AI systems meant to predict social outcomes, like job performance or recidivism. "These problems are hard because we can’t predict the future," Narayanan writes. "That should be common sense. But we seem to have decided to suspend common sense when AI is involved."

This blowback's spark was a 2017 research project from MIT's Joy Buolamwini. She found that major facial recognition systems struggled to identify female and darker-toned faces.

The study and its follow-ups prompted companies to try "fixing" the problem by increasing their systems' accuracy across faces — mostly by gathering more data from underrepresented groups, sometimes in shady ways.

But Buolamwini and others argue that the technology shouldn't be used without being tightly regulated. 

What's next: Companies are tightening access to their AI algorithms, invoking intellectual property protections to avoid sharing details about how their systems arrive at critical decisions.

"The real problem is we citizens have no power to even examine or scrutinize these algorithms," says O'Neil. "They're being used by private actors for commercial gain."

A tug-of-war over biased AI

AI reformers are divided: Some want to fix the algorithms, others aim to keep humans making decisions.

Illustration of a computer with a frowning face

Artificial intelligence

Technology

Automation and AI

The idea that AI can replicate or amplify human prejudice, once argued mostly at the field&#x27;s fringes, has been thoroughly absorbed into its mainstream: <a class="gtmContentClick" data-vars-link-text="Every" data-vars-click-url="https://ai.google/principles/" data-vars-content-id="c7bf3397-a870-4152-9395-83b6bf1e6a67" data-vars-headline="A tug-of-war over biased AI" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://ai.google/principles/" target="_blank">Every</a> <a class="gtmContentClick" data-vars-link-text="major" data-vars-click-url="https://engineering.fb.com/ml-applications/ai-at-f8-2018-open-frameworks-and-responsible-development/" data-vars-content-id="c7bf3397-a870-4152-9395-83b6bf1e6a67" data-vars-headline="A tug-of-war over biased AI" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://engineering.fb.com/ml-applications/ai-at-f8-2018-open-frameworks-and-responsible-development/" target="_blank">major</a> <a class="gtmContentClick" data-vars-link-text="tech" data-vars-click-url="https://www.microsoft.com/en-us/ai/our-approach-to-ai" data-vars-content-id="c7bf3397-a870-4152-9395-83b6bf1e6a67" data-vars-headline="A tug-of-war over biased AI" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://www.microsoft.com/en-us/ai/our-approach-to-ai" target="_blank">tech</a> <a class="gtmContentClick" data-vars-link-text="company" data-vars-click-url="https://www.ibm.com/blogs/research/2019/01/ethically-aligned-ai/" data-vars-content-id="c7bf3397-a870-4152-9395-83b6bf1e6a67" data-vars-headline="A tug-of-war over biased AI" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://www.ibm.com/blogs/research/2019/01/ethically-aligned-ai/" target="_blank">company</a> now makes the necessary noise about &quot;AI ethics.&quot;Yes, but: A critical split divides AI reformers. On one side are the bias-fixers, who believe the systems can be purged of prejudice with a bit more math. (Big Tech is largely in this camp.) On the other side are the bias-blockers, who argue that AI has no place at all in some high-stakes decisions.

From the Valley to D.C., Big Tech players like Facebook, Google and Amazon are under more scrutiny than ever as new technology develops and privacy and antitrust concerns grow in lockstep with companies’ ambitions.

Tech companies appear to be bowing to new privacy rules springing up in Europe, California and elsewhere, putting in place processes to show they're complying.

Yes, but: Some of these moves are smokescreens that allow the companies to avoid making big, painful changes, some legal experts argue — enabled by a legal system that offloads enforcement onto the very companies being regulated.

The big picture: Companies are painting over existing practices with a veneer of rule-following, argues NYU law professor Ari Waldman in an upcoming article for the Washington University Law Review.

"Mere symbols of compliance are standing in for real privacy protections," he writes.

Companies that are meant to be constrained by privacy law are able to "recast and reframe it to benefit themselves," Waldman tells Axios.

The stand-ins, according to Waldman, include privacy policies, impact assessments, trainings, audits and paper trails.

"These things have all the trappings of systems but instead are really just window dressing," he says.

In surveys and interviews with privacy professionals, Waldman turned up a check-the-boxes approach to privacy.

What's happening: As privacy laws in Europe and California kick in, companies are setting up new internal structures to comply with them, says Dominique Shelton Leipzig, a privacy attorney at Perkins Coie.

The other side: "To conclude that assessments aren't working, I think, is a false conclusion," says Al Gidari, a longtime privacy lawyer now at the Stanford Center for Internet and Society.

"Those processes work really well in companies because if they don't, people go to jail, employees get fired, companies get prosecuted," he tells Axios. But it's up to companies to prioritize privacy and implement effective systems.

Gidari argues that internal assessments are necessary at big tech companies like Google, which he represented when it was investigated by the Federal Trade Commission in 2011. It's not possible to formally audit dozens of products and services on a regular basis, he says.

The bottom line: The offloading of enforcement to companies is a result of vague, toothless laws and weakened agencies like the FTC that would otherwise be in charge of enforcement.

"Procedure is not enough," says Waldman. Laws should require a substantive change like a ban on sharing certain data, rather than a process like assessments of whether or not the data is being dealt with correctly.

And penalties should be much higher for wrongdoing, Gidari argues. When the FTC fined Facebook $5B for a privacy violation earlier this year, the company's stock went up. "It's awfully hard to see how that alone is sufficient," Gidari says.

"When you have companies setting the rules, my biggest concern is that it's just going to be streamlined toward the most efficient process for them — but not necessarily the most efficient process for users or the fairest process for users," says Frank Pasquale, a law professor at the University of Maryland.

Go deeper: The global shortage of privacy experts

The privacy smokescreen

Processes can be smokescreens that allow companies to avoid making big, painful changes, some experts argue.

Privacy

Business

Data privacy

Tech companies appear to be bowing to new privacy rules springing up in Europe, California and elsewhere, putting in place processes to show they&#x27;re complying.Yes, but: Some of these moves are smokescreens that allow the companies to avoid making big, painful changes, some legal experts argue — enabled by a legal system that offloads enforcement onto the very companies being regulated.

International students outnumber homegrown talent two to one among newly graduated AI experts, driving American leadership in the critical and increasingly crowded field.

Why it matters: Experts worry that U.S. hostility to immigration is choking this vital pipeline, potentially handing an advantage to competitors like China.

The big picture: Bright minds are the primary fuel for AI advances.

Experts attribute America's primacy in large part to its ability to attract the world's top talent, train them in top universities, and then employ them in academia or the booming private sector.

But as other countries' AI capabilities improve quickly, the balance may be upset.

A landmark new report on AI talent argues that "…any short-term increase in other states’ relative attractiveness — even if counteracted after the fact — can have long-term and potentially irreversible consequences."

Driving the news: In the new report, shared first with Axios, Remco Zwetsloot of the Center for Security and Emerging Technology at Georgetown analyzes the detailed education and work histories of thousands of recent AI graduates in the U.S.

Zwetsloot and his co-authors find that 80% of international students stay in the U.S. after graduation, a proportion that's stayed steady for years. But several looming changes threaten this trend.

Professional opportunities, immigration rules and personal considerations drive international students' decisions to stay or return after school — and immigration is only getting more restricted.

One particular danger, according to the report, is a rollback of Optional Practical Training, a program that allows graduates to work in the U.S. for three years after finishing school.

The program is being challenged in court, and the Trump Administration has prioritized cutting it back.

"This would be a really bad development for the U.S. from an AI competitiveness perspective," Zwetsloot tells Axios.

What's next: The White House has proposed reallocating the number of visas to accommodate more high-skilled immigrants. But Zwetsloot says that's not enough. 

"Ultimately, increasing numerical caps will do a lot more for U.S. retention of the best and brightest than tweaking the allocation system for an insufficient number of slots," he says.

A threat to American AI talent

A new study tracks how the U.S.' anti-immigration climate could choke off its pipeline of AI experts.

International students outnumber homegrown talent two to one among newly graduated AI experts, driving American leadership in the critical and increasingly crowded field.Why it matters: Experts worry that U.S. hostility to immigration is choking this vital pipeline, potentially handing an advantage to competitors like China.

Weapons that down threatening drones — by scrambling their electronics or just plain shooting them out of the sky — are flooding the market, even though most are still illegal in the U.S.

What's new: Just in the last year, hundreds of new products were released, in a scramble to head off an urgent unsolved menace. But off-the-shelf drones are evolving apace, threatening to make a thorny problem even worse.

The big picture: As I wrote this summer, plenty of roadblocks still lie ahead for the counter-drone industry. Fundamentally, many anti-drone systems don't work well — and even if they did, most are illegal in the U.S., except if used by federal agencies.

Driving the news: A new report from the Center for the Study of the Drone at Bard College is a comprehensive census of counter-drone technology.

Altogether, Bard researchers found 537 systems for sale — hundreds more than they found in last year's sweep.

More than 350 of these products are billed for intercepting and disabling drones; the rest simply detect them.

Radio jamming is the most popular method for taking down drones. But other creative approaches involve lasers, nets or even a "sacrificial collision drone."

The report raises two new problems. One is the limited range of many detection systems.

"The response time for successfully shooting down drones is incredibly short if the drone is even moderately fast," says the report's author, Arthur Holland Michel.

Even with a 1 km detection range — which may seem far — several steps remain after an incoming drone is detected: a second check, a decision to intercept, a scramble to ready the relevant weapon…

"By that time, the drone is right over your head," Michel says. "You don't hear this discussed in the marketing materials."

The second problem is the rapid progress of consumer drones, which is creating a "vicious feedback loop," Michel says. Advances that make the devices safer can also make them impervious to some counter-drone systems.

They're fast, with some more expensive drones able to reach 180 mph, or accelerate from 0 to 90 in a second.

They're autonomous. Skydio's newest drone can follow a moving target without human guidance, avoiding obstacles as it goes.

They can fly without GPS. This makes them less prone to dropped signals — but at the same time less susceptible to jamming.

And soon, they'll group into swarms, so that one pilot can fly a horde of drones, opening new doors for drone attacks that are harder to defend against.

The bottom line: "There's nothing on the horizon that will cut the line on this [cycle]," says Michel. "There's nothing that just ends the game. … Until there is, it's going to be like this: a game of cat and mouse."

The explosion of counter-drone technology

Consumer drone advances are keeping the makers of counter-drone weapons scrambling.

A crouching man in fatigues points a rifle-like object toward the sky

A French soldier with an anti-drone rifle. Photo: Chesnot/Getty

Drones

Weapons that down threatening drones — by scrambling their electronics or just plain shooting them out of the sky — are flooding the market, even though most are still illegal in the U.S.What&#x27;s new: Just in the last year, hundreds of new products were released, in a scramble to head off an urgent unsolved menace. But off-the-shelf drones are evolving apace, threatening to make a thorny problem even worse.

Science

Axios Science

Returning to a technology largely discarded since the 1960s, scientists are betting on analog computing to wean AI systems off the monstrous amounts of electricity they currently require.

Why it matters: AI is on track to use up a tenth of the world's electricity by 2025, by one estimate. Cutting back on this consumption has huge climate implications — plus it’s essential for mobile devices and autonomous cars to do complex calculations on the fly.

The background: Analog computing was dethroned by today's dominant digital machines in the 1960s. Since then, computing has been about "higher speed, higher precision, higher throughput," says IBM's Jeff Welser. That's where digital tech shines.

But as AI becomes omnipresent, some of those core requirements of computers are being reconsidered.

A realization is dawning in some corners of the tech world that "maybe we were too quick to dispense with analog 60 years ago," says Eli Yablonovitch, a professor at Berkeley.

What's happening: The neural networks that drive most AI systems rely on multiplying numbers really, really fast. They currently use the precision and power of digital computing for the job. But AI computations may not need to be so precise.

"When you start getting pushed to the limits of what [digital computing] can offer, when you have a new class of problems, then it becomes interesting to revisit analog," says Shahin Farshchi, a computer scientist and VC at Lux Capital.

IBM, several startups, academic researchers and others are doing just that. 

How it works: In a digital computer, everything runs on 1s and 0s — a universal, highly exact human-made language.

But an analog computer is built on the physical properties of its components. It can perform multiplication, for example, by utilizing the properties of transistors.

“The idea is to let the natural dynamics of the physical system solve the problem,” says Garrett Kenyon of the Los Alamos National Laboratory.

These systems come with obstacles: They can be inconsistent and difficult to program, Kenyon says.

Modern experiments with analog technology likely won’t result in a completely analog computer but a hybrid, with an analog portion that approximates an answer that can be fed into a digital part for refinement.

The big picture: There’s a broader resurgence of interest in new and forgotten approaches to computing. 

"Both of the most futuristic areas we're looking at are actually not all digital," Welser says of analog and quantum computing.

Researchers at Los Alamos and elsewhere are developing neuromorphic chips, a subset of analog computing that more closely mirrors neurons in the brain.

"We use ideas regarding the principle of analog computing from the old days, but had to invent completely different ways of implementing them on a modern silicon chip, and had to come up with some completely new ideas as well," says Columbia University's Yannis Tsividis, whose lab is designing hybrid technologies for scientific computing.

What’s next: Analog computing is vying to be a part of the AI explosion. "AI is obviously already a very, very huge thing," says Yablonovitch. "If analog is contributing to that, then it means it has come back after 60 years in the wilderness."

Editor's note: This story has been updated to correct the spelling of Jeff Welser's surname. 

Decades-old analog ideas could buoy modern AI

They could help wean AI off the monstrous amounts of electricity it currently requires.

Illustration: Aïda Amer/Axios. Photos: Authenticated News/Getty Staff, GraphicaArtis/Getty Contributor

Returning to a technology largely discarded since the 1960s, scientists are betting on analog computing to wean AI systems off the monstrous amounts of electricity they currently require.Why it matters: AI is on track to <a class="gtmContentClick" data-vars-link-text="use up a tenth of the world&#x27;s electricity by 2025" data-vars-click-url="https://www.technologyreview.com/s/614005/ai-computing-cloud-computing-microchips/" data-vars-content-id="fcb49e0a-1063-4578-801b-0e6e3e4a96b7" data-vars-headline="Decades-old analog ideas could buoy modern AI" data-vars-event-category="story" data-vars-sub-category="story" data-vars-item="in_content_link" href="https://www.technologyreview.com/s/614005/ai-computing-cloud-computing-microchips/" target="_blank">use up a tenth of the world&#x27;s electricity by 2025</a>, by one estimate. Cutting back on this consumption has huge climate implications — plus it’s essential for mobile devices and autonomous cars to do complex calculations on the fly.

Kaveh Waddell

Stories

The privacy smokescreen

Decades-old analog ideas could buoy modern AI

A threat to American AI talent

The explosion of counter-drone technology

A tug-of-war over biased AI

Newsletter editions

Axios Future: Analog computing for AI — Privacy smokescreens — SF ban’s side effects

Axios Future: Struggle over biased AI — The counter-drone problem — Tapping beer

Axios Future: Obstacle course for AI — Computer co-writers — NYC's many languages

Axios Future: Banning Chinese drones — Future of work today — Finding ancient glyphs

Axios Future: AI washing — Big Beer goes flat — Fun detox