In her words, neuroscientist Poppy Crum, PhD, who will give the opening keynote at AIHce EXP 2023, has spent her career developing technologies that “democratize experiences.” Specifically, she specializes in “bridging technology that captures and represents our senses and the human experience,” she said. This work is significant because human perception and experience are so varied—no two people’s bodies interact with a given technology in exactly the same way. For Crum, studying the physiology of perception, how a person’s perception is impacted by their environment, and how successful they are in that environment has profoundly influenced her work in building technologies that support everyone.

When building technologies, “we need to think not just about what it means for one demographic or building ‘one size fits all’ solutions,” she said. “We need to instead recognize that if we build ‘one size fits all,’ it’s going to impact a lot of people in very different ways.”

Technology and the Human Body

For a relatively low-tech example, amplitude-modulated (AM) radio transmission was a new, rapidly growing technology in the early 20th century. As the technology became more widespread, the radio broadcasting industry would standardize the “voiceband,” the range of frequencies believed to be the bare minimum for transmission of intelligible human speech, between 300 and 3,400 Hertz (Hz). In fact, the speech of the average adult man can be intelligibly transmitted in the 300–3,400 Hz range (at the time, men occupied most radio broadcasting positions). But women tend to have shorter vocal tracts than men. Therefore, they often speak at higher frequencies. Capping a signal at 3,400 Hz distorted the speech of women broadcasters, causing their voices to sound piercing and harsh. One hundred years later, the assumptions embodied in the standardized voiceband still impact telecommunications industries.

One can easily draw parallels between radio in the 1920s and technologies that have proliferated society within recent memory, such as television, the internet, cell phones, machine learning, and artificial intelligence. Often, when a new technology enters widespread use, the technology, its societal implications, or regulatory actions affecting it don’t impact all users equally. To Crum, historical lessons such as the one offered by the development of radio and the voiceband should lead researchers to examine “how we can adapt to the information we have to make better decisions and build better technology—because we should, now that we can.”

“Today, we know the importance of supporting multiple demographics and having wideband audio,” she continued, referring to a broader range of audio frequencies. But it’s important to ensure that new solutions support the diversity of human experiences. For every type of technology being developed today, Crum believes there are opportunities to “better understand the cognitive load or the potential downstream impact it might be having.”

“Rather than building a technology and defining it by how it works,” she said, “instead, we [can] define how it needs to work and figure out what the variables are that we need to introduce and tune for every individual or every demographic and group in different ways.”

How Technology Rewires the Brain

Crum’s expertise in auditory perception has led her to work with Next Sense Inc., a company that designs in-ear wearable sensors, and Dolby Laboratories, an audio technology developer. She holds patents in a range of areas, including augmented sensory systems, technology personalization, and algorithmic transformations and optimizations. As an adjunct professor at Stanford, her work focuses on perception, behavior, the impact of technology on the brain, and how to use this information to build better technology. “I think there’s often a fear of integrating technology because it might change us,” said Crum. “Well, everything we do changes us.”

She explains the neurological effects of technologic change to her students at Stanford using the idea of the cortical homunculus, a representation common in undergraduate psychology and neuroscience textbooks that maps areas and proportions of the human brain to the sensory and motor functions of the body. Parts of the body that are more sensitive or used more often are represented as larger on the homunculus. If a change in technology corresponds to a change in behavior—for example, increased use of the thumbs for typing caused by widespread adoption of smartphones—the homunculus’ relevant body parts will “grow.” In reality, the corresponding parts of the brain are becoming more specialized for the new behavior; the homunculus is only a convenient tool for communicating this information visually.

From this basis, Crum teaches her students to build technology while considering “how it’s going to shape the brain, how it’s going to shape society, and how it’s going to shape interactions,” she said. Her students aim to give people more positive individual experiences through technology—specifically, through video games. “We use video games as a proxy,” she explained, “mainly because it’s a really engaging environment that you can build fairly fast and think about what constraints you’re forcing on the brain and what you’re getting people to engage in. And you can extend that thinking to design for all technologies.” One group of students’ goal was to increase empathy in online gaming communities. They designed a first-person shooter game that required players to trust their opponents in order to succeed.

Creating Better Standards

Crum’s work also crosses over into consumer protection and occupational health, as when she co-chaired the Consumer Technology Association committee that wrote ANSI/CTA 2051, Personal Sound Amplification Performance Criteria. This standard provided the basis for federal regulations for over-the-counter hearing aids propagated by the Food and Drug Administration in 2022.

Writing standards fits with Crum’s other work as “it’s about leveraging technology to help improve people’s success in their environments,” she explained. ANSI/CTA 2051 was crafted “very strategically,” she said, “to help the FDA identify how they could achieve objective quality assessment with product groups that were so heterogeneous in their performance.” That is, every hearing aid user hears differently, experiences hearing loss in a unique way, and has a unique experience with the technology, too. Moreover, researchers are constantly learning more about the factors influencing hearing loss—scientists now know that hearing is not only affected by noise exposure but also by a person’s ingestion of ototoxic drugs, for example. Crum asserts that standards relating to technology use should be written (or rewritten) to reflect new understanding of users’ individual experiences.

“When we define standards, and when we define technologies that interact with the human system, we need to be vigilant in considering if those [standards and technologies] are still the best we can do,” she said. “The goal should always be to understand if we are introducing bias or if we’re truly representing a democratized experience of that technology.”

Technology for More Supportive Solutions

Recent technological developments, including sensors, machine learning, and artificial intelligence, can enable these personalized solutions by allowing occupational and environmental health and safety professionals to collect richer data from better sources at greater resolution and frequency than in the past. OEHS professionals can now obtain dynamic information about both the safety successes and issues within a work environment and use this information to make individual workers most successful, instead of providing the minimum of health and safety support. This abundance of data provides opportunities to rethink ways of supporting people in their environments and to reconsider health and safety standards.

For example, a thermostat might use machine learning and AI to learn a building’s occupancy patterns and automatically adjust its status to save the owners money and resources for heating or cooling the building. But, according to Crum, this thermostat still isn’t “closing the loop” to provide personalized support to the occupants. “It doesn’t know if I’m hot or cold, and it doesn’t know what I’m trying to achieve at that point in time,” she said. The time of day or night and the size, gender, and physical state of a building occupant all have implications for how the thermostat technology may best support their needs. “All of that information we can know,” said Crum, “and we can close the loop to drive those inputs for a device like a thermostat.”

Although Crum cautioned that such technology must also protect people’s privacy, she noted its potential to shift the focus of occupational health away from optimizing the building to using technology to support all the building occupants as equally and effectively as possible. “We can think very differently about what it means for technology to actually be supportive, to understand where there are differences across demographics, which can be quite substantial,” she said. That is, technological solutions, standards, and regulations may be reconfigured to support individuals’ needs within their unique environments, instead of targeting the socially determined “average” person—who, as illustrated in the case of AM radio and voiceband, may be far from representative of everyone.

Poppy Crum will give the opening keynote at AIHce EXP 2023 on Monday, May 22, 2023, from 8 to 9:30 a.m., Mountain time. AIHce EXP 2023 will be held May 22–24 both virtually and onsite at Phoenix Convention Center in Phoenix, Arizona. To learn more about the keynote sessions, view the conference program, or to register, visit the conference website.