If you want a career in AI, learn Python

No, AI isn’t going to take your job. Not yet anyway. As I’ve written, the best uses of artificial intelligence and machine learning (AI/ML) complement human creativity rather than supplant it. Ironically, the best large language models, or LLMs, are trained, perhaps not always legally, using the copyrighted products of human creativity. People and robots are going to peacefully coexist for the foreseeable future.

Even so, some industries are more aggressively embracing AI than others, as revealed in the newest 2022 AI Index Report from Stanford’s Institute for Human-Centered Artificial Intelligence. During the past year, virtually every industry has increased its investments in AI-savvy people, with even higher AI-centric job postings from companies in the following industries: information (5.3%); professional, scientific, and technical services (4.1%); and finance and insurance (3.3%). If you’re worried about your job or simply want to capitalize on this trend, I have one word for you: Python.

Until 2014, academia was the center of the ML universe. No more. Big business has led the AI/ML charge since 2014, and in 2022, businesses released 32 ML models while academia released just three. Academic institutions can’t keep pace with the data, CPU cycles, and money that industry brings.

How much money? Well, while an LLM like GPT-2 cost $50,000 to train back in 2019, PaLM cost roughly $8 million to train, with 360 times more parameters than GPT-2 (which, of course, was cutting edge for its time). Governments could afford this kind of investment, but governments have mostly been concerned with trying (unsuccessfully) to regulate LLMs, so industry has filled the void.

In so doing, businesses’ appetite for AI/ML talent has increased across nearly every American industrial sector. On average, the number of AI/ML-related job postings has ballooned from 1.7% in 2021 to 1.9% in 2022. That number may seem small, but those percentages are of all U.S. job postings. To approach 2% is huge, given how unproven AI/ML remains for most businesses. As I mentioned earlier, some industries have much higher rates of job postings that require AI/ML expertise.

Jobs aren’t the only measure of investment, of course, and in terms of cash, medical and healthcare lead the way with $6.1 billion in AI investments in 2022. Just behind healthcare comes data management, processing, and cloud ($5.9 billion); then fintech ($5.5 billion). These industries make sense, given how those AI funds are being spent. According to the report, businesses use AI in a variety of ways, but the primary areas include robotic process automation (39%), computer vision (34%), natural language text understanding (33%), and virtual agents (33%). As for use cases, the main one embraced in 2022 was service operations optimization (24%). Other popular ones were the creation of new AI-based products (20%), customer segmentation (19%), customer service analytics (19%), and new AI-based enhancement of products (19%).

What does this mean for your job? According to a different study conducted by researchers at the University of Pennsylvania and funded by OpenAI, “around 80% of the U.S. workforce could have at least 10% of their work tasks affected by the introduction of LLMs, while approximately 19% of workers may see at least 50% of their tasks impacted.” Who’s at risk? Accountants, mathematicians, interpreters, creative writers, and more. Who’s not? Those focused on more physical labor, such as cooks, mechanics, or oil-and-gas roustabouts. (Electric vehicles may be coming for that latter group, however.)

This news needn’t be bad, of course. As we’re seeing with software development, AI can remove some of the tedium of a given job while freeing up employees (in this case, developers) to focus on higher-value tasks. For those looking to bolster their chances in this AI-driven future, the Stanford report singles out one particular technology above others: Python.

Python’s impact on data science shouldn’t be a surprise. As I wrote back in 2021, “the language most likely to dominate [data science] is the one that is most accessible to the broadest population within the enterprise.” A year later, this was still true: “As organizations look to a more diverse group to help with data science, Python’s mass appeal makes for an easy on-ramp.” More and more, Python is the lingua franca for experts and novices alike as they dive into data science.

In the Stanford report, Python stands out both for its relative growth compared to other desired skills, but also due to its absolute growth:

There are a number of reasons Python keeps rising to the top for data science, generally, and AI/ML, specifically. Python helps reduce the complexity inherent in AI/ML by providing a bevy of powerful libraries that simplify development. It’s also simple and consistent, with a clear syntax that is human-readable, lowering the bar for becoming proficient with it. Python also comes with a broad, welcoming community to help developers become productive faster, while running across pretty much any platform you might want to use.

Yes, AI might make parts of your job obsolete, given a machine’s ability to do things more efficiently than a human. However, especially for those who pick up Python, there should be plenty of opportunities to embrace the rise of the robotic revolution and extend it to meet your needs (and those of your employer) using Python and other tools.