Once a week.
Subscribe to our weekly newsletter.
Six ways machine learning threatens social justice
Machine learning is a powerful and imperfect tool that should not go unmonitored.
- When you harness the power and potential of machine learning, there are also some drastic downsides that you've got to manage.
- Deploying machine learning, you face the risk that it be discriminatory, biased, inequitable, exploitative, or opaque.
- In this article, I cover six ways that machine learning threatens social justice and reach an incisive conclusion: The remedy is to take on machine learning standardization as a form of social activism.
When you harness the power and potential of machine learning, there are also some drastic downsides that you've got to manage. Deploying machine learning, you face the risk that it be discriminatory, biased, inequitable, exploitative, or opaque. In this article, I cover six ways that machine learning threatens social justice and reach an incisive conclusion: The remedy is to take on machine learning standardization as a form of social activism.
When you use machine learning, you aren't just optimizing models and streamlining business. You're governing. In essence, the models embody policies that control access to opportunities and resources for many people. They drive consequential decisions as to whom to investigate, incarcerate, set up on a date, or medicate – or to whom to grant a loan, insurance coverage, housing, or a job.
For the same reason that machine learning is valuable—that it drives operational decisions more effectively—it also wields power in the impact it has on millions of individuals' lives. Threats to social justice arise when that impact is detrimental, when models systematically limit the opportunities of underprivileged or protected groups.
Here are six ways machine learning threatens social justice
Credit: metamorworks via Shutterstock
1) Blatantly discriminatory models are predictive models that base decisions partly or entirely on a protected class. Protected classes include race, religion, national origin, gender, gender identity, sexual orientation, pregnancy, and disability status. By taking one of these characteristics as an input, the model's outputs – and the decisions driven by the model – are based at least in part on membership in a protected class. Although models rarely do so directly, there is precedent and support for doing so.
This would mean that a model could explicitly hinder, for example, black defendants for being black. So, imagine sitting across from a person being evaluated for a job, a loan, or even parole. When they ask you how the decision process works, you inform them, "For one thing, our algorithm penalized your score by seven points because you're black." This may sound shocking and sensationalistic, but I'm only literally describing what the model would do, mechanically, if race were permitted as a model input.
2) Machine bias. Even when protected classes are not provided as a direct model input, we find, in some cases, that model predictions are still inequitable. This is because other variables end up serving as proxies to protected classes. This is a bit complicated, since it turns out that models that are fair in one sense are unfair in another.
For example, some crime risk models succeed in flagging both black and white defendants with equal precision – each flag tells the same probabilistic story, regardless of race – and yet the models falsely flag black defendants more often than white ones. A crime-risk model called COMPAS, which is sold to law enforcement across the US, falsely flags white defendants at a rate of 23.5%, and Black defendants at 44.9%. In other words, black defendants who don't deserve it are erroneously flagged almost twice as much as white defendants who don't deserve it.
3) Inferring sensitive attributes—predicting pregnancy and beyond. Machine learning predicts sensitive information about individuals, such as sexual orientation, whether they're pregnant, whether they'll quit their job, and whether they're going to die. Researchers have shown that it is possible to predict race based on Facebook likes. These predictive models deliver dynamite.
In a particularly extraordinary case, officials in China use facial recognition to identify and track the Uighurs, a minority ethnic group systematically oppressed by the government. This is the first known case of a government using machine learning to profile by ethnicity. One Chinese start-up valued at more than $1 billion said its software could recognize "sensitive groups of people." It's website said, "If originally one Uighur lives in a neighborhood, and within 20 days six Uighurs appear, it immediately sends alarms" to law enforcement.
4) A lack of transparency. A computer can keep you in jail, or deny you a job, a loan, insurance coverage, or housing – and yet you cannot face your accuser. The predictive models generated by machine learning to drive these weighty decisions are generally kept locked up as a secret, unavailable for audit, inspection, or interrogation. Such models, inaccessible to the public, perpetrate a lack of due process and a lack of accountability.
Two ethical standards oppose this shrouding of electronically-assisted decisions: 1) model transparency, the standard that predictive models be accessible, inspectable, and understandable. And 2) the right to explanation, the standard that consequential decisions that are driven or informed by a predictive model are always held up to that standard of transparency. Meeting those standards would mean, for example, that a defendant be told which factors contributed to their crime risk score -- which aspects of their background, circumstances, or past behavior caused the defendant to be penalized. This would provide the defendant the opportunity to respond accordingly, establishing context, explanations, or perspective on these factors.
5) Predatory micro-targeting. Powerlessness begets powerlessness – and that cycle can magnify for consumers when machine learning increases the efficiency of activities designed to maximize profit for companies. Improving the micro-targeting of marketing and the predictive pricing of insurance and credit can magnify the cycle of poverty. For example, highly-targeted ads are more adept than ever at exploiting vulnerable consumers and separating them from their money.
And insurance pricing can lead to the same result. With insurance, the name of the game is to charge more for those at higher risk. Left unchecked, this process can quickly slip into predatory pricing. For example, a churn model may find that elderly policyholders don't tend to shop around and defect to better offers, so there's less of an incentive to keep their policy premiums in check. And pricing premiums based on other life factors also contributes to a cycle of poverty. For example, individuals with poor credit ratings are charged more for car insurance. In fact, a low credit score can increase your premium more than an at-fault car accident.
6) The coded gaze. If a group of people is underrepresented in the data from which the machine learns, the resulting model won't work as well for members of that group. This results in exclusionary experiences and discriminatory practices. This phenomenon can occur for both facial image processing and speech recognition.
Recourse: Establish machine learning standards as a form of social activism
To address these problems, take on machine learning standardization as a form of social activism. We must establish standards that go beyond nice-sounding yet vague platitudes such as "be fair", "avoid bias", and "ensure accountability". Without being precisely defined, these catch phrases are subjective and do little to guide concrete action. Unfortunately, such broad language is fairly common among the principles released by many companies. In so doing, companies protect their public image more than they protect the public.
People involved in initiatives to deploy machine learning have a powerful, influential voice. These relatively small numbers of people mold and set the trajectory for systems that automatically dictate the rights and resources that great numbers of consumers and citizens gain access to.
Famed machine learning leader and educator Andrew Ng drove it home: "AI is a superpower that enables a small team to affect a huge number of people's lives... Make sure the work you do leaves society better off."
And Allan Sammy, Director, Data Science and Audit Analytics at Canada Post, clarified the level of responsibility: "A decision made by an organization's analytic model is a decision made by that entity's senior management team."
Implementing ethical data science is as important as ensuring a self-driving car knows when to put on the breaks.
Establishing well-formed ethical standards for machine learning will be an intensive, ongoing process. For more, watch this short video, in which I provide some specifics meant to kick-start the process.
Eric Siegel, Ph.D., is a leading consultant and former Columbia University professor who makes machine learning understandable and captivating. He is the founder of the long-running Predictive Analytics World and the Deep Learning World conference series and the instructor of the end-to-end, business-oriented Coursera specialization Machine learning for Everyone. Stay in touch with Eric on Twitter @predictanalytic.
- Prejudice AI? Machine Learning Can Pick up Society's Biases - Big ... ›
- Your safety depends on machine learning - Big Think ›
- Dawn of a new era: AI, machine learning, and robotics - Big Think ›
Geologists discover a rhythm to major geologic events.
- It appears that Earth has a geologic "pulse," with clusters of major events occurring every 27.5 million years.
- Working with the most accurate dating methods available, the authors of the study constructed a new history of the last 260 million years.
- Exactly why these cycles occur remains unknown, but there are some interesting theories.
Our hearts beat at a resting rate of 60 to 100 beats per minute. Lots of other things pulse, too. The colors we see and the pitches we hear, for example, are due to the different wave frequencies ("pulses") of light and sound waves.
Now, a study in the journal Geoscience Frontiers finds that Earth itself has a pulse, with one "beat" every 27.5 million years. That's the rate at which major geological events have been occurring as far back as geologists can tell.
A planetary calendar has 10 dates in red
Credit: Jagoush / Adobe Stock
According to lead author and geologist Michael Rampino of New York University's Department of Biology, "Many geologists believe that geological events are random over time. But our study provides statistical evidence for a common cycle, suggesting that these geologic events are correlated and not random."
The new study is not the first time that there's been a suggestion of a planetary geologic cycle, but it's only with recent refinements in radioisotopic dating techniques that there's evidence supporting the theory. The authors of the study collected the latest, best dating for 89 known geologic events over the last 260 million years:
- 29 sea level fluctuations
- 12 marine extinctions
- 9 land-based extinctions
- 10 periods of low ocean oxygenation
- 13 gigantic flood basalt volcanic eruptions
- 8 changes in the rate of seafloor spread
- 8 times there were global pulsations in interplate magmatism
The dates provided the scientists a new timetable of Earth's geologic history.
Tick, tick, boom
Credit: New York University
Putting all the events together, the scientists performed a series of statistical analyses that revealed that events tend to cluster around 10 different dates, with peak activity occurring every 27.5 million years. Between the ten busy periods, the number of events dropped sharply, approaching zero.
Perhaps the most fascinating question that remains unanswered for now is exactly why this is happening. The authors of the study suggest two possibilities:
"The correlations and cyclicity seen in the geologic episodes may be entirely a function of global internal Earth dynamics affecting global tectonics and climate, but similar cycles in the Earth's orbit in the Solar System and in the Galaxy might be pacing these events. Whatever the origins of these cyclical episodes, their occurrences support the case for a largely periodic, coordinated, and intermittently catastrophic geologic record, which is quite different from the views held by most geologists."
Assuming the researchers' calculations are at least roughly correct — the authors note that different statistical formulas may result in further refinement of their conclusions — there's no need to worry that we're about to be thumped by another planetary heartbeat. The last occurred some seven million years ago, meaning the next won't happen for about another 20 million years.
Long before Alexandria became the center of Egyptian trade, there was Thônis-Heracleion. But then it sank.
Before Alexander the Great established Alexandria around 331 BCE, one of Egypt's primary ports on the Mediterranean Sea between the sixth and fourth centuries BCE was a place called Thônis-Heracleion.
Now researchers from the European Institute for Underwater Archaeology (IEASM), the same organization that first found the city in 2001, have announced the discovery of a couple of fascinating items from the city's heyday. Pinned beneath fallen temple stones is a surprisingly intact Egyptian military vessel from the second century BCE, and researchers have excavated a large cemetery from the fourth century BCE.
Thônis-Heracleion was one of the two primary access points to ancient Egypt from the Mediterranean. (The other, Canopus, was discovered in 1999.) For millennia, experts assumed Thônis-Heracleion were two different lost cities, but it's now known that Thônis is simply the city's Egyptian name, while Heracleion is its Greek name.
Thônis-Heracleion had been the stuff of legend before it was located, mentioned only in rare ancient texts and stone inscriptions. Herodotus seems to have been referring to Thônis-Heracleion's temple of Amun as the place where Heracles first arrived in Egypt. He also described a visit there by Helen with her lover Paris just before the outbreak of the Trojan War. In addition, 400 years later, geographer Strabo wrote that Heraclion, containing the temple of Heracles, had been located opposite Canopus across a branch of the Nile. Today we know Thônis-Heracleion's location as Egypt's Abu Qir Bay. The sunken port is about 6.5 kilometers from the coast and lies beneath ten meters of water.
Both Thônis-Heracleion and Canopus were wealthy in their day, and the temple was an important religious center. This all ended when the Egyptian dynasty created by Ptolemy set out to establish Alexandria as Egypt's center. Thônis-Heracleion and Canopus' trade — and thus wealth — was diverted to the new capital.
It was perhaps just as well, given that natural forces eventually destroyed Thônis-Heracleion. Located on the Mediterranean, the ground upon which it was built became saturated and eventually began to destabilize and liquefy. The temple of Amun probably collapsed around 140 BCE. A series of earthquakes sealed the cty's' fate around 800 CE, sending a 100 square-kilometer chunk of the Nile delta on which it was constructed under the waves. The Mediterranean's rising sea level over the next couple thousand years completed the drowning of Thônis-Heracleion.
Researchers have recovered a large collection of Thônis-Heracleion's treasures revealing an economically rich culture. Coins, bronze statuettes, and over 700 ancient ship anchors have been pulled from the waters. Divers have also identified over 70 shipwrecks. A giant statue of the Nile god Hapi took two and a half years to bring up.
An ancient vessel and a cemetery
Gold mask found in a submerged Greek cemetery.Credit: Egyptian Ministry of Tourism and Antiques
The newly discovered ship was found beneath 16 feet of hard clay, "thanks to cutting-edge prototype sub-bottom profiler electronic equipment," says Ayman Ashmawy of the Egyptian Ministry of Tourism and Antiques.
The military vessel had been moored in Thônis-Heracleion when the temple of Amun collapsed. The temple's enormous blocks fell onto the ship, sinking it. The boat is a rare find — only one other ship of its period has been found. As underwater archaeologist Franck Goddio, one of the scientists who found the city, puts it, "Finds of fast ships from this age are extremely rare."
At 80 feet long, the boat is six times as long as it is wide. Like its dually-named city, it's an amalgam of Greek and Egyptian ship-building techniques. According to expert Ehab Fahmy, head of the Central Department of Underwater Antiquities at IEASM, the boat has some classical construction features such as mortar and tenon joints. On the other hand, it was built to be rowed, and some of its wood was reused lumber, signature traits of Egyptian boat design. Its flat bottom suggests it was built for navigating the shallows of the Nile delta where the river flows into the Mediterranean.
Also found alongside the city's submerged northeastern entrance canal was a large Greek cemetery. The funerary is adorned with opulent remembrances, including a mask made of gold, shown above. A statement by the Egyptian Ministry of Tourism and Antiques describes its significance, as reported by Reuters:
"This discovery beautifully illustrates the presence of the Greek merchants who lived in that city. They built their own sanctuaries close to the huge temple of Amun. Those were destroyed simultaneously and their remains are found mixed with those of the Egyptian temple."
Excavation is ongoing, with more of Egypt's ancient history no doubt waiting beneath the waves.
We are likely to see the first humans walk on Mars this decade.
- Space agencies have successfully sent three spacecraft to Mars this year.
- The independent missions occurred at around the same time because Earth and Mars were particularly close to each other last summer, providing an opportune time to launch.
- SpaceX says it hopes to send a crewed mission to Mars by 2026, while the U.S. and China aim to land humans on the planet in the 2030s.
Spacecraft from three of the world's space agencies reached Mars this year.
In February, the United Arab Emirates' Hope space probe entered the Martian orbit, where it is studying the planet's weather cycles. That same month, NASA's Perseverance rover touched down on Mars, where it will soon begin collecting rock samples that could contain signs of ancient life. And in May, China successfully landed its Zhurong rover on the Martian surface, becoming the second nation to ever do so.
All three missions launched in the summer of 2020. The timing was no coincidence: once every two years, Earth and Mars come especially close together because their orbits are "at opposition," which is when the Earth-Mars distance is smallest during the 780-day synodic period. It is an opportune window to send spacecraft to Mars.
The handful of spacecraft currently exploring the Martian surface and atmosphere are scheduled to conduct their experiments for periods ranging from months to years. Some even plan to collect materials to return to Earth. For example, NASA's Perseverance will store its rock samples in protective tubes and leave them behind for a smaller "fetch rover" to pick up on a future mission.
Photo of Martian surface taken by the Perseverance roverNASA/JPL-Caltech
If all goes well, an Airbus spacecraft dubbed the Earth Return Orbiter (ERO) will carry the samples back to Earth in 2031. It would be the first time a space mission has returned Martian matter to Earth. But before the decade's end, space agencies have some other missions that aim to study the Red Planet.
Europe & Russia
NASA is not the only space agency aiming to find evidence of life on the Red Planet. In 2023, Roscosmos and the European Space Agency plan to land their Rosalind Franklin rover on the Martian surface, where it will drill into rock and analyze soil composition for signs of past — or possibly present — alien life.
The joint mission is part of a long-term Mars project that began in 2016. This second phase was initially planned for 2020, but due in part to the COVID-19 pandemic, the space agencies decided to postpone the launch to 2022.
"We want to make ourselves 100% sure of a successful mission. We cannot allow ourselves any margin of error. More verification activities will ensure a safe trip and the best scientific results on Mars," said ESA Director General Jan Wörner.
In 2022, the Japanese Aerospace Exploration Agency (JAXA) plans to send to Mars its TEREX lander, which will "precisely measure the amount of water molecules and oxygen molecules, and search for water resources and the possibility of life on Mars," JAXA wrote.
In 2024, JAXA also plans to launch a uniquely bold interplanetary mission that will involve sending a probe to orbit Mars, landing on the Martian moon Phobos, collecting surface samples, and then returning those samples to Earth in 2029. JAXA says the mission has two main objectives: (1) to investigate whether the Martian moons are captured asteroids or fragments that coalesced after a giant impact with Mars; and (2) to clarify the mechanisms controlling the surface evolution of the Martian moons and Mars.
Following the successful landing of its Zhurong rover this year, China released a roadmap of its plans for additional Mars voyages. The first is an uncrewed mission scheduled for 2030, with crewed missions planned for 2033, 2035, 2037, and 2041. As the International Space Station project is coming to a close, China is in the process of building its own space station; earlier this year it launched into orbit the first part of its station, which will take 10 more missions to assemble.
Elon Musk's California-based aerospace company has its sights on two Mars voyages: a cargo-only mission in 2022 and a human mission by 2026. The crewed mission would involve building a propellant depot and preparing a site for future crewed flights. Getting to Mars will first require an orbital test of SpaceX's Starship rocket, which the company hopes to conduct this year.
Regarding the long-term future of humans on the Red planet, Musk once told Ars Technica:
"I'll probably be long dead before Mars becomes self-sustaining. But I'd like to at least be around to see a bunch of ships land on Mars."
In 2014, the Indian Space Research Organization executed its first interplanetary trip with its Mars Orbiter Mission. It marked the first time an Asian nation reached Martian orbit and also the first time a nation successfully reached the Red planet on its maiden voyage. India has plans for a follow-up Mars Orbiter Mission 2, but it remains unclear when that will occur and what the mission will entail.
In February, the chief of the Indian Space Research Organisation said the nation would only launch a Mars mission after Chandrayaan-3, India's upcoming mission to the Moon, which is expected to launch in 2022.