It’s not all doom and gloom.
With all the headlines warning about how A.I. is poised to take our jobs, steal our likeness for nefarious purposes, and even upend democracy, it’s understandable to be skeptical (and maybe even a little afraid) of this new technology. But while there are plenty of valid reasons to be wary of this new technology and want to implement protections, there are also a lot of good things that are happening because of artificial intelligence. Here are some of the benefits of using A.I.
What are the benefits of using A.I.?
There are advantages and disadvantages to using artificial intelligence. The big benefit of A.I. that came up a lot when discussing ChatGPT is the idea that it can help humans do certain tasks more efficiently and without certain kinds of user error. (It’s important to note, though, that even ChatGPT will admit it can generate incorrect or biased responses — I asked the program to tell me its benefits, and it warned, “It’s important to note that ChatGPT, like any language model, has limitations. It may generate incorrect or biased responses, struggle with nuanced or ambiguous queries, and lack real-time awareness of current events beyond its knowledge cutoff date.”) A number of companies have integrated A.I. into their customer service, which helps customers get answers to some of their common questions even if live representatives aren’t available. It can also help with certain rote or repetitive tasks, like data entry, or transcription. And it turns out, there are a lot of good things A.I. has already done across a range of industries that go beyond speeding up tedious written tasks.
How A.I. is being used to detect cancer
You might think medicine would be the one industry that’s immune from the impacts of A.I., but this technology is already being used in healthcare for a variety of purposes, and a huge one is detecting cancer. A study released last week by Harvard Medical School found that A.I. can help neurosurgeons treat brain tumors, or gliomas — an umbrella term that encompasses the most common type of brain tumor. Typically, removing a patient’s tumor without damaging the brain tissue requires doctors to send a tissue sample to a pathologist to analyze certain factors, such as what type of cancer the patient has and whether they cut the right tissue. This usually takes 10 to 15 minutes while the patient is on the surgical table — with their skull open. The study found that new machine learning may reduce the amount of time it takes to analyze that tissue sample and do so more accurately, which would decrease the amount of time surgery takes. The A.I. could also assess the tissue more accurately — the tool, called CHARM, had 93 percent accuracy at identifying tumors and classifying them into the three major types of gliomas.
A.I. is not just being used to identify and treat brain cancer. A.I. may also be able to help treat breast cancer by decreasing the time it takes for radiologists to read mammograms, so they can look at more over a given period of time. That could mean that some people’s cancer is spotted quicker. And in Hungary, some hospitals and clinics are experimenting with A.I. software that flags potential cancer cells on a mammogram. So far, it’s in its early stages, but the technology has already spotted some cancerous growths that human radiologists have missed. More research is needed on these tools before they’re widely implemented, but the results seem promising so far.
Other ways A.I. is being used in healthcare
There are lots of other ways this technology is being used in healthcare settings besides cancer detection. For instance, machine-learning programs are helping doctors more quickly identify people who may be in danger for sepsis — which contributes to one-third of hospital deaths in the U.S. and can become deadly in a matter of hours. One program looked at patients’ electronic health records for risk factors of sepsis and scanned their current lab tests and vital signs to identify those who were at high risk of developing septic shock. A trial of the program suggested that it was helpful in diagnosing sepsis early and reducing mortality.
A.I. is also being used to help diagnose certain complications from diabetes, heart disease, and many other conditions.
Tackling climate change
How can robots stop the rising global temperatures? The answer is a little abstract, but a number of projects that use artificial intelligence have emerged to combat climate change. One such effort is called This Climate Does Not Exist, which uses deepfakes and A.I.-generated images — but for good. These pictures depict how any address will look if it’s hit with wildfires, smog, or floods. In that way, it could help motivate people to get involved to end climate change by seeing for themselves how it can impact their hometown. Microsoft is also building an A.I. for Earth program that aggregates data from NASA, the European Space Agency, and other government agencies around the world to work as a “geospatial search engine” that can help speed up climate-related decisions and “avert environmental disaster.” How exactly that will work remains to be seen just yet, but it shows that there’s a lot of potential for this technology to be used in positive ways.
Combating world hunger
A.I.’s ability to analyze a lot of information quickly and make predictions is at the core of how it can solve complex global issues like world hunger. For that issue specifically, it can use predictive models to pinpoint areas with nutrient deficiencies. For instance, the Nutrition Early Warning Sign program looks at data from satellite imagery, rainfall, temperature, and vegetation health to predict where there could be threats to crops, which would then give humans the opportunity to intervene before those threats come to be. Some researchers are using the technology to identify areas where people may have micronutrient deficiencies, which usually requires expensive and/or time consuming lab tests to detect. This could, down the line, allow for public health interventions to happen sooner.
How A.I. could predict the next pandemic
You know how every few months or so, scientists identify a new Covid variant of concern? That’s done by tracking the variants and their mutations, which is something A.I. can help with. One computational model analyzed more than 6.4 million Covid sequences to identify patterns within the mutations and predict which variant would become the dominant one. The model correctly predicted the dominant variants in December 2021 (Omicron) and March 2022 (BA.2). While the program can’t determine which mutations are likely to arise in the virus, it is able to predict which ones are most likely to become widespread, which can help us plan for future variants. (That said, there are many areas of the pandemic that A.I. theoretically could have changed the game, but failed to.)
For better or for worse, artificial intelligence is probably going to change a lot of aspects of the way we live our lives. While many of these programs and advancements are in the very early stages, there’s a lot of potential for the technology to have positive impacts.
