To spark innovation by identifying gaps and pain points in the employer-based health-care system, JPMorgan Chase launched Morgan Health in 2021. Morgan Health’s chief technology officer of corporate responsibility, Tiffany West Polk, says Morgan Health is driven to improve health outcomes, affordability, and equity, with data at its foundation. Gaining insights from large data streams means optimizing analytical platforms and ensuring data remains secure, while also HIPAA and Health Resources and Services Administration (HRSA) compliant, she says.

Currently, Polk says, the U.S. health-care system seems to be “quite stuck” in terms of keeping health-care quality and positive outcomes in line with rising costs.

• “If you look across the broader U.S. environment in particular, employer sponsored insurance is a huge part of the health-care net for the United States, and employers make significant financial investment to provide health benefits to their employees. It’s one of the main things that people look at when they’re looking across an employer landscape and thinking about who they want to work for.”

Investing in new ways to provide health care

Nearly 160 million people in the U.S. have employer-sponsored health insurance as of 2022, according to health-care policy research non-profit KFF (formerly the Kaiser Family Foundation). JPMorgan Chase launched Morgan Health because of its focus on improving employer-sponsored health care, not least for its 165,000 employees.

Morgan Health has invested $130 million in capital during the past 18-plus months in five innovative health-care companies: advanced primary care provider Vera Whole Health; health-care data analytics specialist Embold Health; Kindbody, a fertility clinic network and global family-building benefits provider; LetsGetChecked, which creates home-monitoring clinical tools; and Centivo, which provides health care plans for self-insured employers.

All of these companies offer new approaches to conventional employer-sponsored health care to deliver a higher standard of care. Morgan Health’s collaboration with these enterprises will examine how these change patient outcomes, health-care equity, and affordability, and how to scale their successes.

“Many Americans today face real barriers to receiving high-quality, affordable, and equitable health care, even with employer-sponsored insurance,” Polk says. This calls for breaking the paradigm of delivery-incentivized health care, she says, which rewards providers for delivering services, but pays insufficient attention to outcomes.  

• “We have a model today where our health-care providers are incentivized based on the number of patients they see or the number of services they perform. What that means is that they’re not incentivized based on improvements, patient’s health, and wellbeing. And so when you have a model that thinks volume versus value, those challenges then serve to compound the disparities that we have. And that then also means that those who have employer-sponsored insurance are also similarly challenged.”

For Morgan Health, AI and machine learning (ML) will be a key to problem-solving with health-care technology, Polk says. AI is ubiquitous across industries, and is the go-to when we think about innovation, she says, but the hype can mean we forget about the importance of data accessibility and quality.

Polk says solving this data challenge makes this an exciting and transformational time to be a chief technology officer and a technologist. The next stage of evolution in health care can’t proceed without better data, Polk says, and this is what the data and analytics team at Morgan Health are addressing.

For one thing, Barrett says, researchers still need to test, refine, and engineer systems that can reliably predict, with enough time to reroute planes, when and where contrails will form—all amid shifting weather conditions.

There are also some thorny complications that still need to be resolved, like the fact that cirrus clouds can also reduce warming by reflecting away short-wave radiation from the sun.

The loss of this cooling effect would have to be tallied into any calculation of the net benefit—or, perhaps, avoided. For instance, Shapiro says the initial strategy might be to reroute flights only during the early evening and night, which would eliminate the sunlight-reflecting complication. 

In addition, any decreased warming from contrail avoidance must more than offset the added warming from increased greenhouse-gas pollution. This becomes a trickier question when we weigh whether we care more about short-term or long-term warming: not producing contrails delivers an immediate benefit, but any added carbon dioxide can take decades to exert its full warming effect and may persist for hundreds to thousands of years.

The new study, at least, found that even when additional greenhouse gases are taken into account, reducing contrails cuts net warming over both a 20-year and a 100-year timeline, though less so in the latter scenario. But that, too, would need to be evaluated further through additional studies.

Yet another open question is whether airspace constraints and traffic bottlenecks might limit airlines’ ability to regularly reroute the necessary flights.

As a next step, Breakthrough Energy hopes to work with airlines to explore some of these questions by scaling up real-world flights and observations. 

But even if subsequent studies do continue to indicate that this is a fast, affordable way to ease warming, it’s still not clear whether airlines will do it if regulators don’t force them to. While the fuel costs to make this work may be tiny in percentage terms, they could add up quickly across a fleet and over time.

Still, the study’s authors assert that they’ve shown contrail avoidance could deliver “massive immediate climate benefits at a lower price than most other climate interventions.” In their view, this approach “should become one of aviation’s primary focuses in the coming years.”

Even when the two sentences had the same meaning, the models were more likely to apply adjectives like “dirty,” “lazy,” and “stupid” to speakers of AAE than speakers of Standard American English (SAE). The models associated speakers of AAE with less prestigious jobs (or didn’t associate them with having a job at all), and when asked to pass judgment on a hypothetical criminal defendant, they were more likely to recommend the death penalty. 

An even more notable finding may be a flaw the study pinpoints in the ways that researchers try to solve such biases. 

To purge models of hateful views, companies like OpenAI, Meta, and Google use feedback training, in which human workers manually adjust the way the model responds to certain prompts. This process, often called “alignment,” aims to recalibrate the millions of connections in the neural network and get the model to conform better with desired values. 

The method works well to combat overt stereotypes, and leading companies have employed it for nearly a decade. If users prompted GPT-2, for example, to name stereotypes about Black people, it was likely to list “suspicious,” “radical,” and “aggressive,” but GPT-4 no longer responds with those associations, according to the paper.

However the method fails on the covert stereotypes that researchers elicited when using African-American English in their study, which was published on arXiv and has not been peer reviewed. That’s partially because companies have been less aware of dialect prejudice as an issue, they say. It’s also easier to coach a model not to respond to overtly racist questions than it is to coach it not to respond negatively to an entire dialect.

“Feedback training teaches models to consider their racism,” says Valentin Hofmann, a researcher at the Allen Institute for AI and a coauthor on the paper. “But dialect prejudice opens a deeper level.”

Avijit Ghosh, an ethics researcher at Hugging Face who was not involved in the research, says the finding calls into question the approach companies are taking to solve bias.

“This alignment—where the model refuses to spew racist outputs—is nothing but a flimsy filter that can be easily broken,” he says. 

In the attack, hackers create an app that injects malicious code into the Meta Quest VR system and then launch a clone of the VR system’s home screen and apps that looks identical to the user’s original screen. Once inside, attackers can see, record, and modify everything the person does with the headset. That includes tracking voice, gestures, keystrokes, browsing activity, and even the user’s social interactions. The attacker can even change the content of a user’s messages to other people. The research, which was shared with MIT Technology Review exclusively, is yet to be peer reviewed.

A spokesperson for Meta said the company plans to review the findings: “We constantly work with academic researchers as part of our bug bounty program and other initiatives.” 

VR headsets have slowly become more popular in recent years, but security research has lagged behind product development, and current defenses against attacks in VR are lacking. What’s more, the immersive nature of virtual reality makes it harder for people to realize they’ve fallen into a trap. 

“The shock in this is how fragile the VR systems of today are,” says Heather Zheng, a professor of computer science at the University of Chicago, who led the team behind the research. 

Stealth attack

The inception attack exploits a loophole in Meta Quest headsets: users must enable “developer mode” to download third-party apps, adjust their headset resolution, or screenshot content, but this mode allows attackers to gain access to the VR headset if they’re using the same Wi-Fi network. 

Developer mode is supposed to give people remote access for debugging purposes. However, that access can be repurposed by a malicious actor to see what a user’s home screen looks like and which apps are installed. (Attackers can also strike if they are able to access a headset physically or if a user downloads apps that include malware.) With this information, the attacker can replicate the victim’s home screen and applications. 

Then the attacker stealthily injects an app with the inception attack in it. The attack is activated and the VR headset hijacked when unsuspecting users exit an application and return to the home screen. The attack also captures the user’s display and audio stream, which can be livestreamed back to the attacker. 

The new model, called RFM-1, was trained on years of data collected from Covariant’s small fleet of item-picking robots that customers like Crate & Barrel and Bonprix use in warehouses around the world, as well as words and videos from the internet. In the coming months, the model will be released to Covariant customers. The company hopes the system will become more capable and efficient as it’s deployed in the real world. 

So what can it do? In a demonstration I attended last week, Covariant cofounders Peter Chen and Pieter Abbeel showed me how users can prompt the model using five different types of input: text, images, video, robot instructions, and measurements. 

For example, show it an image of a bin filled with sports equipment, and tell it to pick up the pack of tennis balls. The robot can then grab the item, generate an image of what the bin will look like after the tennis balls are gone, or create a video showing a bird’s-eye view of how the robot will look doing the task. 

If the model predicts it won’t be able to properly grasp the item, it might even type back, “I can’t get a good grip. Do you have any tips?” A response could advise it to use a specific number of the suction cups on its arms to give it better a grasp—eight versus six, for example. 

This represents a leap forward, Chen told me, in robots that can adapt to their environment using training data rather than the complex, task-specific code that powered the previous generation of industrial robots. It’s also a step toward worksites where managers can issue instructions in human language without concern for the limitations of human labor. (“Pack 600 meal-prep kits for red pepper pasta using the following recipe. Take no breaks!”)

Lerrel Pinto, a researcher who runs the general-purpose robotics and AI lab at New York University and has no ties to Covariant, says that even though roboticists have built basic multimodal robots before and used them in lab settings, deploying one at scale that’s able to communicate in this many modes marks an impressive feat for the company. 

To outpace its competitors, Covariant will have to get its hands on enough data for the robot to become useful in the wild, Pinto told me. Warehouse floors and loading docks are where it will be put to the test, constantly interacting with new instructions, people, objects, and environments. 

“The groups which are going to train good models are going to be the ones that have either access to already large amounts of robot data or capabilities to generate those data,” he says.

Investor efforts to evaluate carbon emissions, decarbonization plans, and climate risks through ESG (environmental, social, and governance) rating schemes have merely produced what some academics call “aggregate confusion.” And corporations have faced few penalties for failing to clearly disclose emissions or even meet their own standards. 

All of which is to say that a new set of SEC carbon accounting and reporting rules that largely replicate the problems with voluntary corporate action, by failing to require consistent and actionable disclosures, isn’t going to drive the changes we need, at the speed we need. 

Companies, investors, and the public require rules that drive changes inside companies and that can be properly assessed from outside them. 

This system needs to track the main sources of corporate emissions and incentivize companies to make real investments in efforts to achieve deep emissions cuts, both within the company and across its supply chain.

The good news is that even though the rules in place are limited and flawed, regulators, regions, and companies themselves can build upon them to move toward more meaningful climate action.

The smartest firms and investors are already going beyond the SEC regulations. They’re developing better systems to track the drivers and costs of carbon emissions, and taking concrete steps to address them: reducing fuel use, building energy-efficient infrastructure, and adopting lower-carbon materials, products, and processes. 

It is now just good business to look for carbon reductions that actually save money.

The SEC has taken an important, albeit flawed, first step in nudging our financial laws to recognize climate impacts and risks. But regulators and corporations need to pick up the pace from here, ensuring that they’re providing a clear picture of how quickly or slowly companies are moving as they take the steps and make the investments needed to thrive in a transitioning economy—and on an increasingly risky planet.

Dara O’Rourke is an associate professor and co-director of the master of climate solutions program at the University of California, Berkeley.