The 11 Most-Read Stories of 2022
As we near the end of the year, revisit the stories and research that grabbed the most attention in 2022. Articles cover our popular AI Index report, a tool to detect autism more quickly, an approach to capture human traffickers, and an assessment of autonomous car black-box safety validation algorithms.
The AI Index is a perennial favorite, covering the state of the industry through a wide lens including education, technology, jobs, investment, diversity, and – this year – ethics. This at-a-glance look at the report’s key infographics offers a good starting place for the in-depth analysis.
In the latest AI Index, we find that private investment doubled since 2022, AI became more affordable/higher performing, U.S. and China dominated cross-country research collaborations, research on fairness and transparency exploded, and AI patents soared. This story offered major highlights of the 2022 report.
Stanford scientists recorded the individual activity of thousands of animal neurons from eight different brain areas over several days while the animals repeatedly performed a visual discrimination task. What emerges is a detailed picture of how the brain processes visual cues – from perception to discrimination to behavioral response – a portrait that may have implications for technologies such as brain-computer interfaces and computer vision.
Stanford scholars developed an algorithm that may determine if someone has autism by looking at brain scans and also predict the severity of autism symptoms in individual patients. With further honing, the algorithm could lead to earlier diagnoses, more targeted therapies, and broadened understanding of autism’s origins in the brain.
What job characteristics lead to higher wages and mobility? Using machine learning, a Stanford scholar showed that the words used in a dataset of more than 1 million online job postings explain 87% of the variation in salaries across a vast proportion of the labor market.
Scholars from Stanford’s Human Trafficking Lab combine AI and satellite imagery to track forced labor at deforestation sites in the Amazon rainforest. Forced labor is hard to find, says one of the scholars, but the satellite imagery can help identify where it’s happening in real time for inspectors to intervene more quickly and effectively.
Autonomous car designers rely on simulations to test their vehicles’ abilities. But how good are these simulations? Stanford researchers surveyed black-box safety validation algorithms to find that out of the nine systems tested, only two provide anything more than falsification validation, just one offers most-likely-failure testing, and another offers probability estimation. There’s room for improvement, they say.
Stanford scholars used natural language processing and machine learning to analyze more than 18,000 research papers that contained results from brain scans gathered primarily using functional magnetic resonance imaging (fMRI) technology. These scans study which parts of the brain leap into action when people are asked to do various things such as speak, move, rest, think logically, remember events from the past, or feel particular emotions. They found different parts of the brain work together in surprising ways that differ from current neuroscientific wisdom. In particular, the study calls into question our current understanding of how brains process emotion. The work could lead to a better understanding of mental disorders and, eventually, to more successful treatments.
In short, yes. A research team examined the gaps between the availability and accessibility of AI-mediated communication tools that enable interpersonal communication assisted by an intelligent agent (think autocomplete in email and texts, or transcription and translation tools). The team found that AI-mediated communication technology is “not a monolith” – categories were not used or experienced equally by all users – and that device and internet access, age, user speech characteristics, and AI tool literacy were barriers to adoption.
Neo-colonialism dominates the Global South, says Stanford fellow Sabelo Mhlambi. By engendering perpetual dependence on products and internet infrastructures, global tech monopolies like Facebook, Amazon, and Google make it almost impossible for people to determine their own futures on their own terms. Here he discusses ways historically marginalized groups can decide and build their own socio-technical futures.
Decision-making algorithms ingest problem-relevant information from the environment and produce an action. Think an AI algorithm that takes in patient vital signs and outputs a diagnosis, or a stock-trading system that synthesizes daily market prices and suggests stock buys. But building an agent in a highly uncertain environment is a challenge for any developer. In a new book, a trio of scholars recommend various approaches for designers solving different kinds of problems.
Stanford HAI's mission is to advance AI research, education, policy, and practice to improve the human condition. Learn more.