10 Weird but Useful New Jobs for AI Technology

AI is currently hated by many, and it’s no wonder why, with the proliferation of artificially created, soulless slop flooding the modern internet. But it’s not AI’s fault that a new generation of unimaginative folks are using AI-generated content to spam social media and post “24-Hour Relax Music Mediterranean” videos on YouTube.

Sadly, this detracts from AI as a revolutionary scientific advance. Therefore, these actually useful uses aim to highlight AI’s technological prowess and unconventional potential—from the practical to the exploratory.

Related: 10 Biggest Misconceptions about Artificial Intelligence

10 Extreme Microbes May Reveal New Antibiotics

When researchers need new antibiotics, they explore fungi, bacteria, or animals to scour their DNA, survival strategies, or the weird ooze they expel. Yet there are crazier lifeforms on Earth that live in pure salt, boiling acid, and scorching deep-sea vents.

These most extreme of extremophiles are microbes belonging to a domain called Archaea. These ancient lifeforms are separate from bacteria and animals on the tree of life, and can thrive where others can’t. Though they look like bacteria, they occupy the craziest environs, including hot springs at Yellowstone and the scorching underwater vents that may have helped Earth’s first lifeforms evolve.

Now, AI is looking at the biological substances and survival mechanics that make, and are made by, Archaea. These biochemicals evolved to resist extreme stresses, so it makes sense that they may lead to new antibiotics or other health innovations that Archaea have quietly been developing for billions of years.^[1]

9 AI Helps Combat New York’s Overdose Crisis

New York has a drug overdose crisis, so researchers are leveraging AI to analyze social work, public health, and data science repositories to refine drug policies and gauge how they work in real-world scenarios.

AI can help address the crisis by analyzing the vast amount of data collected through the multi-disciplinary HEALing Communities Study (HCS), the largest-ever NIH-funded initiative to combat overdoses. AI can help by discovering patterns to predict where interventions are needed, how to increase their effectiveness, and even which individuals might be most at risk.

AI can help divert necessary resources like naloxone, an opioid-reversal drug that can avert overdoses, to where they’re most needed. AI also helps craft public health messages and can then collect and analyze real-time data to see how well these strategies work across 16 targeted New York communities.^[2]

8 Spiders’ Visual Tricks Even Fool Computers

Some jumping spiders have developed an incredible evolutionary scare tactic: their backs look like wasps or praying mantises.

To determine just how accurately (and objectively) this illusion works, researchers used AI to try and differentiate between jumping spiders, wasps, and praying mantises. In other words, are these illusions so good that they could fool a computer? Or is it just our human perceptual abilities that make the illusions appear so dang convincing?

It turns out the illusions really are good. When an algorithm tried its digital hand at differentiating between jumping spiders, wasps, and praying mantises, it failed about 20% of the time.

Oddly enough, the Maratus vespa spider imitates a wasp even in its courtship display, which is a wild way to try to attract a mate. It may do so to gain a female’s attention before breaking the deception and reassuring her by offering, “Hey, I’m actually a nice guy.”^[3]

7 Meet the Digital Nurses of the Future

NVIDIA and Hippocratic AI are exploring the efficacy of AI nurses, which can answer low-risk calls. And none too late because populations are booming, health is dropping, and medical professionals are overwhelmed, with a significant shortage of healthcare workers forecast by 2030.

This leaves a big gap for potential digital nurses. The NVIDIA model is designed through collaborations with hospitals and can assess around 400 diseases before referring patients to the proper portals. For patients, being able to get immediate attention (even if from a digital entity) can be a major boon in an era where many feel like they’re being ignored.

Of course, the AI will not make diagnoses and will not replace human nurses. It’s more of a telehealth avenue to ameliorate worker shortages and assist patients with follow-up care or simple tips—perhaps clearing up a prescribed drug regimen question or other such query… for a modest hourly fee.^[4]

6 Robots Are Writing Scientific Papers

Writing a huge scientific paper is a chore, even for the scientists excited to share the new scimitar-toothed tyrannosaur they just discovered (scientists hate writing essays just as much as you do).

Accordingly, AI usage is increasing in scientific journals. Scientists studying AI use in prolific science publications analyzed well over 1 million papers in their study, including those from arXiv and Nature. They found that AI use in computer science papers increased by 22% between 2020 and 2024. Mathematics papers, however, saw a much smaller AI use increase of 9%.

Are scientists getting lazier? Probably not. Rather, this trend may be due to the increased globalization of science as well as publishing access. A portion of AI increases occurred in countries with fewer English speakers, who could have used large language models (LLMs) to help make their academic papers more accessible and relatable to the world at large.^[5]

5 Chatbots That Never Sleep

Customer service queries aren’t always complicated, and people often grow angry with slow service. That’s becoming a problem of the past. AI chatbots like ChatGPT are becoming more intuitive and in tune with diverse databases and web searches. As a result, chatbot customer service agents are increasingly serving as standalone tools or assistants for human support agents.

Businesses of varying sizes can utilize chatbots when customers require quick, simple answers. Or at the other end of the service spectrum, when customers have complex queries or request obscure data that would take too long to find or compile manually. Chatbots can also work 24/7 for increased accessibility. However, customers will no doubt always find reasons to complain.^[6]

4 Drive-Thru Bots vs. Burritos

Taco Bell is “learning a lot” by trialing AI in its drive-thrus. However, “learning a lot” isn’t exactly a comforting sentence when you’re hungry for potato burritos at 3 a.m.

Taco Bell debuted the technology at more than 500 locations in the U.S.. Still, it may be rethinking the strategy after various, uh, glitches. In one instance, the AI kept asking a customer to order a drink to go along with their Mountain Dew, leading to understandable frustration. In another incident, a man reportedly crashed the system by ordering 18,000 waters.

A similar trial at McDonald’s led to bacon being added to sundaes and an accidental hundred-dollar chicken nugget order.

Therefore, despite being introduced to reduce mistakes (because we all know AI never makes mistakes) and speed up drive-through ordering, it seems that AI is breeding a novel type of disgruntled customer. But hey, there’s no bad publicity, right?^[7]

3 AI Cracks the Code of Double Stars

From the pitch to the pitch-blackness of space, AI is prized for its versatility. New AI models can help detect and analyze binary stars, one of the building blocks of the universe.

Roughly half of Sun-like stars come in pairs, making binary stars an essential cosmological building block that gives birth to exosolar systems, star clusters, galaxies, and galaxy clusters.

Orbiting stars also affect one another in many ways, including gravity, magnetic fields, and radiation. So ascertaining their features is insanely challenging, requiring tens of millions of parameters and hundreds of hours of computing time—even networks may take weeks to resolve a single binary star system. As a result, relatively few systems have precisely measured properties.

The solution? Use new AI models, trained on datasets that predict stellar properties, and then analyze existing observations to find matches. Such AI models can reach up to about 99% accuracy while being orders of magnitude faster. What took a week on a supercomputer can now take minutes on a laptop.^[8]

2 Smarter Offsides: AI in Sports Tactics

AI is being used in all facets of life, even sports, to offer insights. Despite its reputation for making certain things worse (art, writing, etc.), AI may have mostly positive effects on sport officiating.

For example, in soccer, it’s used for Semi-Automated Offside Technology (SAOT) by tracking players’ positions in real time. Viewers know the potential frustration (or elation) brought about by this somewhat recent system. Still, it aims to get calls right more often.

Additionally, researchers spent more than three years developing models that can digitally track all the players on the field at every moment to detail their actions. Perfecting such systems, which are trained on real games, such as those of the 2022 FIFA World Cup, can help bridge AI accessibility and soccer to evolve the game by informing tactics and improving on-field calls, like those made by Video Assistant Referees (VARs).^[9]

1 Chatbots Could Talk to Aliens for Us

Chatbots could be our ambassadors to the universe. Sure, finding simple alien life would be an existential discovery, but finding intelligent life could be absolutely transformative (though hopefully they won’t want to destroy us). But could we ever hope to communicate with smart aliens, even a bit? Perhaps, with the help of AI.

An unconventional scientific idea proposes that we send a large language model (LLM) into space. Instead of just sending out music, math, or messages, as we’ve done, we should basically beam a chatbot into space, via laser or other means, for the benefit of potential alien civilizations.

In this way, simplified of course for ease of explainability, any existing ETs could converse with this chatbot. Through such conversations, they can learn human languages, receive replies representative of humanity, and ask the bot questions about ourselves—but hopefully not about our weaknesses or KFC’s secret recipe (coincidentally, one of our weaknesses).^[10]