In an initial study, Canary analyzed speech recordings with AI and identified early stage Alzheimer’s with 96 percent accuracy.
Dave Arnold retired in his 60s and began spending time volunteering in local schools. But then he started misplacing items, forgetting appointments and losing his sense of direction. Eventually he was diagnosed with early stage Alzheimer’s.
“Hearing the diagnosis made me very emotional and tearful,” he said. “I immediately thought of all my mom had experienced.” His mother suffered with the condition for years before passing away. Over the last year, Arnold has worked for the Alzheimer’s Association as one of its early stage advisors, sharing his insights to help others in the initial stages of the disease.
Arnold was diagnosed sooner than many others. It's important to find out early, when interventions can make the most difference. One promising avenue is looking at how people talk. Research has shown that Alzheimer’s affects a part of the brain that controls speech, resulting in small changes before people show other signs of the disease.
Now, Canary Speech, a company based in Utah, is using AI to examine elements like the pitch of a person’s voice and their pauses. In an initial study, Canary analyzed speech recordings with AI and identified early stage Alzheimer’s with 96 percent accuracy.
Developing the AI model
Canary Speech’s CEO, Henry O’Connell, met cofounder Jeff Adams about 40 years before they started the company. Back when they first crossed paths, they were both living in Bethesda, Maryland; O’Connell was a research fellow at the National Institutes of Health studying rare neurological diseases, while Adams was working to decode spy messages. Later on, Adams would specialize in building mathematical models to analyze speech and sound as a team leader in developing Amazon's Alexa.
It wasn't until 2015 that they decided to make use of the fit between their backgrounds. ““We established Canary Speech in 2017 to build a product that could be used in multiple languages in clinical environments,” O'Connell says.
The need is growing. About 55 million people worldwide currently live with Alzheimer’s, a number that is expected to double by 2050. Some scientists think the disease results from a buildup of plaque in the brain. It causes mild memory loss at first and, over time, this issue get worse while other symptoms, such as disorientation and hallucinations, can develop. Treatment to manage the disease is more effective in the earlier stages, but detection is difficult since mild symptoms are often attributed to the normal aging process.
O’Connell and Adams specialize in the complex ways that Alzheimer’s effects how people speak. Using AI, their mathematical model analyzes 15 million data points every minute, focusing on certain features of speech such as pitch, pauses and elongation of words. It also pays attention to how the vibrations of vocal cords change in different stages of the disease.
To create their model, the team used a type of machine learning called deep neural nets, which looks at multiple layers of data - in this case, the multiple features of a person’s speech patterns.
“Deep neural nets allow us to look at much, much larger data sets built out of millions of elements,” O’Connell explained. “Through machine learning and AI, we’ve identified features that are very sensitive to an Alzheimer’s patient versus [people without the disease] and also very sensitive to mild cognitive impairment, early stage and moderate Alzheimer's.” Based on their learnings, Canary is able to classify the disease stage very quickly, O’Connell said.
“When we’re listening to sublanguage elements, we’re really analyzing the direct result of changes in the brain in the physical body,” O’Connell said. “The brain controls your vocal cords: how fast they vibrate, the expansion of them, the contraction.” These factors, along with where people put their tongues when talking, function subconsciously and result in subtle changes in the sounds of speech.
Further testing is needed
In an initial trial, Canary analyzed speech recordings from phone calls to a large U.S. health insurer. They looked at the audio recordings of 651 policyholders who had early stage Alzheimer’s and 1018 who did not have the condition, aiming for a representative sample of age, gender and race. They used this data to create their first diagnostic model and found that it was 96 percent accurate in identifying Alzheimer’s.
Christian Herff, an assistant professor of neuroscience at Maastricht University in the Netherlands, praised this approach while adding that further testing is needed to assess its effectiveness.
“I think the general idea of identifying increased risk for cognitive impairment based on speech characteristics is very feasible, particularly when change in a user’s voice is monitored, for example, by recording speech every year,” Herff said. He noted that this can only be a first indication, not a full diagnosis. The accuracy still needs to be validated in studies that follows individuals over a period of time, he said.
Toby Walsh, a professor of artificial intelligence at the University of New South Wales, also thinks Canary’s tool has potential but highlights that Canary could diagnose some people who don’t really have the disease. “This is an interesting and promising application of AI,” he said, “but these tools need to be used carefully. Imagine the anxiety of being misdiagnosed with Alzheimer’s.”
As with many other AI tools, privacy and bias are additional issues to monitor closely, Walsh said.
Other languages
A related issue is that not everyone is fluent in English. Mahnaz Arvaneh, a senior lecturer in automatic control and systems engineering at the University of Sheffield, said this could be a blind spot.
“The system may not be very accurate for those who have English as their second language as their speaking patterns would be different, and any issue might be because of language deficiency rather than cognitive issues,” Arvaneh said.
The team is expanding to multiple languages starting with Japanese and Spanish. The elements of the model that make up the algorithm are very similar, but they need to be validated and retrained in a different language, which will require access to more data.
Recently, Canary analyzed the phone calls of 233 Japanese patients who had mild cognitive impairment and 704 healthy people. Using an English model they were able to identify the Japanese patients who had mild cognitive impairment with 78 percent accuracy. They also developed a model in Japanese that was 45 percent accurate, and they’re continuing to train it with more data.
The future
Canary is using their model to look at other diseases like Huntington’s and Parkinson’s. They’re also collaborating with pharmaceuticals to validate potential therapies for Alzheimer’s. By looking at speech patterns over time, Canary can get an indication of how well these drugs are working.
Dave Arnold and his wife dance at his nephew’s wedding in Rochester, New York, ten years ago, before his Alzheimer's diagnosis.
Dave Arnold
Ultimately, they want to integrate their tool into everyday life. “We want it to be used in a smartphone, or a teleconference call so that individuals could be examined in their home,” O’Connell said. “We could follow them over time and work with clinical teams and hospitals to improve the evaluation of patients and contribute towards an accurate diagnosis.”
Arnold, the patient with early stage Alzheimer’s, sees great promise. “The process of getting a diagnosis is already filled with so much anxiety,” he said. “Anything that can be done to make it easier and less stressful would be a good thing, as long as it’s proven accurate.”
Sharing land and other resources among farmers isn’t new. But research shows it may be increasingly relevant in a time of climatic upheaval.
Sharing land and other resources among farmers isn’t new. But research shows it may be increasingly relevant in a time of climatic upheaval, potentially influencing “farmers to adopt environmentally friendly practices and agricultural innovation,” according to a 2021 paper in the Journal of Economic Surveys. Farmers might share knowledge about reducing pesticide use, says Heather Frambach, a supply chain consultant who works with farmers in California and elsewhere. As a group they may better qualify for grants to monitor soil and water quality.
Most research around such practices applies to cooperatives, whose owner-members equally share governance and profits. But a collective like Carman Ranch’s — spearheaded by fourth-generation rancher Cory Carman, who purchases beef from eight other ranchers to sell under one “regeneratively” certified brand — shows when producers band together, they can achieve eco-benefits that would be elusive if they worked alone.
Vitamins and minerals in soil pass into plants through their roots, then into cattle as they graze, then back around as the cows walk around pooping.
Carman knows from experience. Taking over her family's land in 2003, she started selling grass-fed beef “because I really wanted to figure out how to not participate in the feedlot world, to have a healthier product. I didn't know how we were going to survive,” she says. Part of her land sits on a degraded portion of Zumwalt Prairie replete with invasive grasses; working to restore it, she thought, “What good does it do to kill myself trying to make this ranch more functional? If you want to make a difference, change has to be more than single entrepreneurs on single pieces of land. It has to happen at a community level.” The seeds of her collective were sown.
Raising 100 percent grass-fed beef requires land that’s got something for cows to graze in every season — which most collective members can’t access individually. So, they move cattle around their various parcels. It’s practical, but it also restores nutrient flows “to the way they used to move, from lowlands and canyons during the winter to higher-up places as the weather gets hot,” Carman says. Meaning, vitamins and minerals in soil pass into plants through their roots, then into cattle as they graze, then back around as the cows walk around pooping.
Cory Carman sells grass-fed beef, which requires land that’s got something for cows to graze in every season.
Courtesy Cory Carman
Each collective member has individual ecological goals: Carman brought in pigs to root out invasive grasses and help natives flourish. Probert also heads a more conventional grain-finished beef collective with 100 members, and their combined 6.5 million ranchland acres were eligible for a grant supporting climate-friendly practices, which compels them to improve soil and water health and biodiversity and make their product “as environmentally friendly as possible,” Probert says. The Washington veg farmer reduced tilling and pesticide use thanks to the ecoservices of visiting cows. Similarly, a conventional hay farmer near Carman has reduced his reliance on fertilizer by letting cattle graze the cover crops he plants on 80 acres.
Additionally, the collective must meet the regenerative standards promised on their label — another way in which they work together to achieve ecological goals. Says David LeZaks, formerly a senior fellow at finance-focused ecology nonprofit Croatan Institute, it’s hard for individual farmers to access monetary assistance. “But it's easier to get financing flowing when you increase the scale with cooperatives or collectives,” he says. “This supports producers in ways that can lead to better outcomes on the landscape.”
New, smaller scale farmers might gain the most from collective and cooperative models.
For example, it can help them minimize waste by using more of an animal, something our frugal ancestors excelled at. Small-scale beef producers normally throw out hides; Thousand Hills’ 50 regenerative beef producers together have enough to sell to Timberland to make carbon-neutral leather. In another example, working collectively resulted in the support of more diverse farms: Meadowlark Community Mill in Wisconsin went from working with one wheat grower, to sourcing from several organic wheat growers marketing flour under one premium brand.
Another example shows how these collaborations can foster greater equity, among other benefits: The Federation of Southern Cooperatives has a mission to support Black farmers as they build community health. It owns several hundred forest acres in Alabama, where it teaches members to steward their own forest land and use it to grow food — one member coop raises goats to graze forest debris and produce milk. Adding the combined acres of member forest land to the Federation’s, the group qualified for a federal conservation grant that will keep this resource available for food production, and community environmental and mental health benefits. “That's the value-add of the collective land-owner structure,” says Dãnia Davy, director of land retention and advocacy.
New, smaller scale farmers might gain the most from collective and cooperative models, says Jordan Treakle, national program coordinator of the National Family Farm Coalition (NFFC). Many of them enter farming specifically to raise healthy food in healthy ways — with organic production, or livestock for soil fertility. With land, equipment and labor prohibitively expensive, farming collectively allows shared costs and risk that buy farmers the time necessary to “build soil fertility and become competitive” in the marketplace, Treakle says. Just keeping them in business is an eco-win; when small farms fail, they tend to get sold for development or absorbed into less-diversified operations, so the effects of their success can “reverberate through the entire local economy.”
Frambach, the supply chain consultant, has been experimenting with what she calls “collaborative crop planning,” where she helps farmers strategize what they’ll plant as a group. “A lot of them grow based on what they hear their neighbor is going to do, and that causes really poor outcomes,” she says. “Nobody replanted cauliflower after the [atmospheric rivers in California] this year and now there's a huge shortage of cauliflower.” A group plan can avoid the under-planting that causes farmers to lose out on revenue.
It helps avoid overplanted crops, too, which small farmers might have to plow under or compost. Larger farmers, conversely, can sell surplus produce into the upcycling market — to Matriark Foods, for example, which turns it into value-add products like pasta sauce for companies like Sysco that supply institutional kitchens at colleges and hospitals. Frambach and Anna Hammond, Matriark’s CEO, want to collectivize smaller farmers so that they can sell to the likes of Matriark and “not lose an incredible amount of income,” Hammond says.
Ultimately, farming is fraught with challenges and even collectivizing doesn’t guarantee that farms will stay in business. But with agriculture accounting for almost 30 percent of greenhouse gas emissions globally, there's an “urgent” need to shift farming practices to more environmentally sustainable models, as well as a “demand in the marketplace for it,” says NFFC’s Treakle. “The growth of cooperative and collective farming can be a huge, huge boon for the ecological integrity of the system.”
How AI changes things
Sure, you would reply, but AI changes everything. With artificial intelligence, the machine itself will develop some sort of autonomy, however ill-defined. It will have a will of its own. And this will, if it reflects anything that seems human, will not be benevolent. With AI, the claim goes, the machine will somehow know what it must do to get rid of us. It will threaten us as a species.
Well, this fear is also not new. Mary Shelley wrote Frankenstein in 1818 to warn us of what science could do if it served the wrong calling. In the case of her novel, Dr. Frankenstein’s call was to win the battle against death — to reverse the course of nature. Granted, any cure of an illness interferes with the normal workings of nature, yet we are justly proud of having developed cures for our ailments, prolonging life and increasing its quality. Science can achieve nothing more noble. What messes things up is when the pursuit of good is confused with that of power. In this distorted scale, the more powerful the better. The ultimate goal is to be as powerful as gods — masters of time, of life and death.
Should countries create a World Mind Organization that controls the technologies that develop AI?
Back to AI, there is no doubt the technology will help us tremendously. We will have better medical diagnostics, better traffic control, better bridge designs, and better pedagogical animations to teach in the classroom and virtually. But we will also have better winnings in the stock market, better war strategies, and better soldiers and remote ways of killing. This grants real power to those who control the best technologies. It increases the take of the winners of wars — those fought with weapons, and those fought with money.
A story as old as civilization
The question is how to move forward. This is where things get interesting and complicated. We hear over and over again that there is an urgent need for safeguards, for controls and legislation to deal with the AI revolution. Great. But if these machines are essentially functioning in a semi-black box of self-teaching neural nets, how exactly are we going to make safeguards that are sure to remain effective? How are we to ensure that the AI, with its unlimited ability to gather data, will not come up with new ways to bypass our safeguards, the same way that people break into safes?
The second question is that of global control. As I wrote before, overseeing new technology is complex. Should countries create a World Mind Organization that controls the technologies that develop AI? If so, how do we organize this planet-wide governing board? Who should be a part of its governing structure? What mechanisms will ensure that governments and private companies do not secretly break the rules, especially when to do so would put the most advanced weapons in the hands of the rule breakers? They will need those, after all, if other actors break the rules as well.
As before, the countries with the best scientists and engineers will have a great advantage. A new international détente will emerge in the molds of the nuclear détente of the Cold War. Again, we will fear destructive technology falling into the wrong hands. This can happen easily. AI machines will not need to be built at an industrial scale, as nuclear capabilities were, and AI-based terrorism will be a force to reckon with.
So here we are, afraid of our own technology all over again.
What is missing from this picture? It continues to illustrate the same destructive pattern of greed and power that has defined so much of our civilization. The failure it shows is moral, and only we can change it. We define civilization by the accumulation of wealth, and this worldview is killing us. The project of civilization we invented has become self-cannibalizing. As long as we do not see this, and we keep on following the same route we have trodden for the past 10,000 years, it will be very hard to legislate the technology to come and to ensure such legislation is followed. Unless, of course, AI helps us become better humans, perhaps by teaching us how stupid we have been for so long. This sounds far-fetched, given who this AI will be serving. But one can always hope.
This article originally appeared on Big Think, home of the brightest minds and biggest ideas of all time.
