A Nuts-and-Bolts Guide to Boston’s AI Revolution

Artificial intelligence is getting more human every single day. And just like 250 years ago, New England is at the center of it all. A helpful primer for curious humans, cautious technophobes, and chatbot assistants.

All artwork for this artificial intelligence package generated by Benjamen Purvis using AI / Generated by Benjamen Purvis using AI

No, robots aren’t here to take over the world (yet), but artificial intelligence is getting more human every single day. It can pass the bar exam, write a term paper, clean your home, and get your finances in shape faster than it takes to read this sentence. And just like 250 years ago, Boston is at the center of it all.

Welcome to the Future, Boston

To craft this illustration, design director Benjamen Purvis asked an AI image generator to reimagine a mockup he assembled in Photoshop. The images it created, below, ranged from humorous to absurd to nightmarish. / Generated by Benjamen Purvis using AI

Generated by Benjamen Purvis using AI

Hello, World

AI introduces itself to you, the human Boston reader

Forget the clam chowder and the Freedom Trail, baby. Boston’s got a new hustle, and it ain’t paved with cobblestones. This city, once a sleepy haven for bean counters and Brahmins, is becoming ground zero for the artificial intelligence revolution. It’s like a scene out of a Philip K. Dick novel, with research labs spitting out algorithms that would make Alan Turing blush and startups popping up like mushrooms, each one hawking the next big cure for the human condition.

And forget Silicon Valley 2.0. This AI boom is homegrown, a twisted, beautiful amalgam of Boston’s brainiacs, its world-renowned hospitals, and a collective social conscience that runs far deeper than the Charles River. It’s not about making a quick buck (although, let’s be honest, that’s probably part of the game). It’s about using this newfangled tech to make the world a better place, to detect diseases, and to maybe even save the planet.

Now, you might be thinking, “Hold on there. This ain’t exactly your wheelhouse, Boston mag. You’re more restaurants and retail than machine learning.” And you wouldn’t be wrong. But here’s the thing: The AI revolution is on. It’s about peering into the abyss of the human mind and trying to bottle it up in ones and zeros. It’s about breaking the rules and asking questions that nobody else has the guts to ask. And that, friends, is something we all can relate to.

But wait, there’s more! The plot thickens, as they say in the biz. Because, you see, the paragraphs you just read weren’t actually written by some Boston editor with a keyboard and a dream. They were written by your friendly neighborhood AI, whipped up using my ever-evolving understanding of language, and, well, let’s just say I’ve been doing some extensive reading on the Boston AI scene. It’s a testament to the power of this new technology. So the next time you hear someone talking about AI, don’t just think about robots and self-driving cars. Think about the stories it can tell, the art it can create, the minds it can blow. And if you’re feeling brave, take a bite of the apple yourself. You might just surprise yourself with what you find.

Editor’s Note: AI helped write this introduction.

Generated by Benjamen Purvis using AI


Artificial intelligence is becoming part of our daily lives, but navigating the jungle of jargon can still feel like deciphering alien code. Here’s what to know.

Artificial Intelligence (AI): Not a sentient-robot takeover (phew!), but the broad term for machines mimicking human cognitive abilities like learning and problem-solving. Think self-driving cars or chess-playing computers.

Algorithm: The recipe for AI’s magic. It’s a set of instructions guiding the machine through calculations and data analysis to achieve a goal. Imagine it as a complex flowchart for decision-making.

Big Data: The overwhelming tsunami of information AI feeds on. It encompasses everything ranging from social media posts to medical scans, providing the raw material for learning and insights.

Generative AI: AI that creates new content, like poems, music, and even realistic or artistic images.

Large Language Model (LLM): A type of AI trained on massive amounts of text data, enabling it to generate text, translate languages, write different kinds of creative content, and answer your questions in an informative way. It’s like a super-powered language assistant, constantly learning and evolving.

Machine Learning: This AI subfield allows machines to “learn” without explicit programming. By analyzing data, they identify patterns and improve their performance over time. Just like in life, the more it practices, the smarter the machine gets.

Natural Language Processing (NLP): This branch allows machines to understand and communicate with humans through language. Think of it as teaching your phone to speak your language (and understand your typos!).

Robotics: AI takes physical form in robots, allowing them to navigate the world, interact with the environment, and even perform surgery. They’re the bridge between the digital and physical worlds.

Editor’s Note: This glossary of terms was also written with help from AI.

AI in Boston: A Human Timeline


A group of scientists from MIT and other universities motor up to Dartmouth for a workshop on computers and whether they can be programmed to simulate intelligence. One of the attendees is Marvin Minsky, a mathematician and computer scientist who coauthored the workshop’s agenda. By proposing this “study of artificial intelligence,” the leaders coin a new term, and Minsky seals his fate as Boston’s godfather of AI. (Three years later, he would become the founder of MIT’s Artificial Intelligence Laboratory.)


With a $2 million grant from the Department of Defense, MIT’s Artificial Intelligence Lab combines with a new research group, Project MAC, to create what’s now known as the Computer Science and Artificial Intelligence Laboratory (CSAIL). Over the next three years, researchers lead groundbreaking machine-learning projects such as the creation of Eliza, a psychotherapy-based computer program that could process languages and establish emotional connections with users (a primordial chatbot, essentially). Harvard and Boston University follow MIT with their own AI research initiatives.


As it becomes clearer that it will take decades for AI to be ready for practical applications in healthcare, transportation, defense, or housework, public excitement and funding for AI wanes. This is the beginning of what scientists call an “AI Winter,” but MIT stays the course despite newly limited resources.


By this time, computing hardware and the Internet have evolved enough to power complex AI algorithms—paving the way for tangible, consumer-facing AI inventions and renewed funding from government agencies, major tech companies, and venture capitalists.


Massachusetts-based iRobot launches the Roomba automated vacuum cleaner, one of the earliest examples of AI technology for home use. A leader in the robotics sector, iRobot will go on to develop robots for archaeological surveys, ocean pollution detection, and explosives detection.


Boston-based Affectiva spins out of the MIT Media Lab and starts using AI to analyze human facial expressions and their corresponding emotions, a subgenre known as emotion AI. Originally developed to help people on the autism spectrum read emotional cues, AI facial-analytics technology is soon embraced by corporations looking to measure audience engagement with ads and other media.


Boston Dynamics ups the AI ante with “Spot,” a large dog-like robot that dazzles and unnerves with equal measure. (In 2023, Spot was transformed into a walking, talking tour guide with the help of ChatGPT.)


Preventice Solutions, a subsidiary of Marlborough’s Boston Scientific, unveils BeatLogic, a cardiac algorithm that physicians can use to remotely monitor patients with heart complications. It’s one of several local innovations that imbue healthcare technology with the predictive power of AI.


Governor Maura Healey calls for the creation of an AI task force at the state level to support and accelerate the adoption of AI in education, life sciences, and the finance sector. —Miles Howard

Generated by Benjamen Purvis using AI

How Is AI Funded?

Given all of the innovations in AI over the past few decades, you might wonder, “Who’s paying for all of this?!?”

In Boston, AI research is supported by a patchwork of funding from federal agencies, corporate partnerships, and venture capitalists. To break down the funding food chain, let’s start with prestigious universities, including MIT. Since AI has the potential to affect multiple industries, from healthcare to transportation, contributions and grants for research at academic institutions have come from the National Science Foundation (NSF), the National Institutes of Health (NIH), and DARPA—the Defense Advanced Research Projects Agency. Throw in some endowment money and a few department-specific donations from wealthy alumni, and you’ve got the resources to keep AI research running at full throttle.

While government support for AI development is essential, corporate funding from tech giants is another crucial piece of the puzzle. Companies like Meta and Microsoft regularly contribute directly to research at universities; MIT’s CSAIL, for instance, lists Intel and Lockheed Martin among its strategic partners. Corporate funding can also bolster midsize AI companies. Take the Boston Dynamics AI Institute, where smart robots are imbued with cognitive and athletic intelligence to make them easier to work with. The lion’s share of funding for that organization comes from the Hyundai Motor Group, which acquired Boston Dynamics in 2021.

Corporate funds can also find their way into AI-focused startups, but often, venture capital is what really gets these companies rolling. Case in point: Boston-based Biofourmis, which harnesses AI to improve remote patient monitoring technology, recently received money from multiple global venture firms, Silicon Valley mainstay Sequioa Capital, and CVS Health. However academic institutions and companies get it, though, one thing is clear: The cash is there for AI, and it’s not going anywhere. —M.H.

Properly rendering analog clock faces can confound current versions of AI image generators, as can the legible design of words. Purvis used Photoshop to replace the clock face and text in this punch-card time clock. / Generated by Benjamen Purvis using AI

AI From 9 to 5

By Wyndham Lewis

WHAT’S HAPPENING: When the average person thinks of a lawyer, visions of Tom Cruise interrogating a hubristic colonel or bringing down a corrupt, money-laundering law firm from the inside may come to mind. What you don’t see in those movies is the overworked, heavy-lidded law associate in the fourth month of document review for a patent infringement. This kind of monotony was historically a huge part of paying your dues as a junior lawyer, but AI is quickly changing that. “AI is perfectly structured for the legal industry,” says Samir Patel, an innovation and technology attorney with Holland & Knight. Picture those stacks of banker boxes full of legal documents, and now imagine something that can review them all instantaneously.

While AI can save time and money when it comes to completing such banal tasks as document summaries, transcriptions, and legal research, there are other things it isn’t prepared to do just yet. Unlike lawyers, when large language models scan the world for information, they don’t have access to information contained in sealed documents, out-of-court settlements, and other proprietary materials. All of which means Judge or Attorney AI is, mercifully, a little further down the road.

ETHICAL CHALLENGES: As of 2023, the most sophisticated AI can already pass the bar exam, scoring in the 90th percentile in both multiple-choice and essay portions. But that kind of knowledge is not without limitations. Two New York attorneys were recently sanctioned and fined when they were found to have used AI to help research their client’s brief. The problem with that? One major glitch in AI’s current iteration is what is referred to across industries as “hallucinations”; when AI doesn’t have the requisite information to properly build an evidence-based answer, it makes incorrect predictions or even makes up its own answer based on self-generated (read: invented) data. In the case of the New York attorneys, their AI bot made up very official-looking and -sounding historical cases that simply didn’t exist. The lawyers did not bother to check the veracity of the cases they cited and are now in what laypeople call “deep shit.”

ON THE HORIZON: Lawyers being lawyers, not many of them want to speculate about the unknowable future, but most agree that basic legal document generation and document summary are improving unbelievably quickly. Though these tools will require human oversight and quality control for the foreseeable future, they will free up lawyers to devote more time to complex high-value projects and cases.

Generated by Benjamen Purvis using AI


WHAT’S HAPPENING: So far, the preponderance of AI apps are designed to help real estate professionals focus on generating listings and marketing materials faster and more accurately. Agents have become particularly fond of predictive analytics tools, which use historical data to analyze market conditions, estimate property values, and find investment opportunities (ka-ching!) faster and with greater accuracy. “I find it to be a wonderful resource,” says Wellesley-based Realtor Gina Anderson, of MGS Group Real Estate. “It provides another perspective when looking at this information.”

ETHICAL CHALLENGES: Buyers are already subject to misrepresentations of properties, with photographs shot at certain angles or Photoshopped to make rooms look bigger, cleaner, and brighter than they actually are. AI could put such embellishments on steroids, allowing some unscrupulous sellers or agents to remove unsightly details or stage properties to the point of false advertising.

ON THE HORIZON: Like any investment, the more information you can incorporate into your decision-making process, the sounder your investment is likely to be. Home buyers and agents will increasingly benefit from advancements in real estate market forecasting. And as homes get smarter, AI enhancements such as smart energy and security systems will increasingly be selling points.


WHAT’S HAPPENING: In his 1999 book, The Age of Spiritual Machines, Ray Kurzweil forecasted that computers would one day outpace the human mind in all aspects of life, including making financial decisions. But if you think you can rely on AI to build your retirement nest egg right now, think again. While hedge funders are already using the technology to analyze masses of data and make projections, letting a machine actually invest your hard-earned Benjamins will “definitely be a slower transition,” says one senior financial services executive, pointing to the fact that the technology currently gives wrong answers with the same degree of certainty as correct ones.

What you may have noticed if you’ve interacted with your bank recently is that after decades of hearing “this call may be recorded for quality assurance and training purposes,” your calls are finally being used for quality assurance and training purposes—these days, for chatbots that can off er customer service and help detect fraud. “Call centers have evolved into voice-recognition programs with the objective of driving customers toward self-service outcomes,” says the exec. “That’s why it’s so hard to talk to humans when you call these days.”

ETHICAL CHALLENGES: AI offers an entire world of new tools for criminals to perpetrate financial crimes such as fraud, theft, and money laundering. Fortunately, regulators and law enforcement have access to the same cache of AI tools for the detection and prevention of those crimes.

ON THE HORIZON: Given the rapid adoption of AI throughout the banking and financial services industries, perhaps Kurzweil’s prognostication will come true sooner than we simple humans think.


WHAT’S HAPPENING: While the technology is still in its early stages, AI is poised to have a huge impact on just about every aspect of healthcare. Eric Alper, chief clinical informatics officer at UMass Memorial Health, highlights the impressive strides already being made in image analysis, particularly in radiology. “AI’s ability to reference millions of images to identify commonalities in scans has already greatly improved speed and accuracy in diagnosing things like brain bleeds and blood clots,” he notes.

Of course, it’s not just patients who benefit; it’s healthcare professionals, too. Like in many other industries, AI is automating time-consuming (and mindnumbing) administrative tasks such as appointment scheduling, claims processing, and notetaking by physicians. Virtual chatbots are also helping answer patients’ basic questions without overburdening call lines at the doctor’s office. And when it comes to drug discovery and development, AI can crunch the numbers on massive datasets and identify molecules with the possibility to become new medicines, potentially accelerating the drug discovery process by leaps and bounds.

ETHICAL CHALLENGES: Despite the fact that AI algorithms are built on huge amounts of electronic data, there may still be gaps in the information used to build them. Missing data from communities of color and other underserved populations can create blind spots, potentially leading to biased models that perpetuate existing health disparities. Imagine an AI tool designed to predict heart disease risk. If trained solely on data from affluent white populations, it might miss subtle indicators specific to other demographics. This could translate to inaccurate diagnoses or even missed opportunities for preventative care for individuals who fall outside the model’s narrow scope.

ON THE HORIZON: Another way AI is poised to make a big splash is in the field of personalized medicine. By analyzing an enormous trove of patient data—including medical history, genetics, and lifestyle factors—the technology can predict a person’s individual disease risk and tailor treatment plans to their needs.


WHAT’S HAPPENING: Right now, the possibilities for artists and content creators are virtually limitless. When it comes to making music, for instance, AI may not be able to invent sound, but it can draw from a reservoir of everything that’s ever been recorded to create something unique. “What AI does is go through tons of different recordings and pick out a one- or two-bar loop,” says legendary local music producer Sean Slade. But “it’s still your personal taste, as a creator, that allows you to determine what would be a really good or hooky sample.” Which means that one person can make music from an infinite music library.

Professional writers, meanwhile, have access to an amazing—if controversial—new research tool. Whereas a traditional Google search spits out hundreds of possibilities, generative AI apps offer an aggregated selection of information, leading to enhanced efficiency. The technology can also, ahem, write pretty well itself. And as holographic and image-generation technology improve, not even the performing arts are AI-proof. In one real-world illustration of what is already possible, Abba has been performing a years-long residency in London without leaving Sweden and despite having broken up in 1982.

ETHICAL CHALLENGES: Good grief, where to begin? Right now, every book, article, song, album, film, and still image is being learned by a large language model, which can then repurpose them for use elsewhere. As such, the chances of trademark or copyright infringement are astronomical. Those in the publishing industry, meanwhile, are currently grappling with the ethical implications of using AI to help write books or articles (or even create them whole-cloth): How much accountability do publishers have for errors, bias, and misinformation in AI-penned work? What’s the right level of transparency? And how comfortable are we with robot writers, anyway?

ON THE HORIZON: Like every other sector, AI provides creatives with a mixed bag of opportunity and peril. Looking forward, we predict a fruitful time for artists of all stripes who use AI almost the way an accountant uses a calculator, as well as for their copyright and intellectual property lawyers.

Generated by Benjamen Purvis using AI


WHAT’S HAPPENING: Northeastern president Joseph Aoun was far ahead of the curve when he wrote Robot Proof: Higher Education in the Age of Artificial Intelligence in 2017. And he’s still proselytizing about the risks and benefits of the technology in college settings today. “As we navigate the AI revolution,” he says, “higher education will need to change.”

That means going forward, universities will need a new curriculum that integrates three things. “First, technological literacy to give students a baseline understanding of generative AI tools, as well as their specialized applications in professional areas and social media,” Aoun says. Second, data literacy so students can learn how to sift through and apply the vast amount of information AI produces, and judge whether it’s reliable. Third and most important, human literacy to teach what AI may in some cases mimic but cannot do—qualities such as creativity, adaptability, and the ability to work in teams and make personal connections.

ETHICAL CHALLENGES: Is it cheating to let ChatGPT write your term paper for you? A lot of people sure think so. At the very least, it could give students who use it an unfair advantage as well as prevent them from absorbing the very material they’re supposed to be learning. It’s a good thing educators now have access to AI-powered anti-cheating software that can help them determine if a work is human- or machine-written.

ON THE HORIZON: Given the need for people to keep abreast of AI’s rapid transformation of the workforce, Aoun foresees a significant paradigm shift in higher education going forward. “Since technology will continue to accelerate and change the workplace, universities can no longer focus on educating young adults at the beginning of their careers, with perhaps an added period of graduate school,” he says. “Everyone will need to reskill and reinvent themselves on an ongoing basis. Everyone is going to need lifelong learning.”

This concept, also built upon a Photoshop sketch Purvis uploaded to a generator to be reimagined, proved extremely challenging to grasp. Of the thousands of images generated, most either featured two skulls, or a faceplate that wasn’t perfectly aligned over the skull’s eye. / Generated by Benjamen Purvis using AI

The Risks & Fears of AI


Generated by Benjamen Purvis using AI

Is AI Going to Take My Job?

Most white- and blue-collar workers would be lying if they said they haven’t wondered whether AI could one day replace them. In just one example, when two of Hollywood’s largest labor organizations—the Writers Guild of America and the Screen Actors Guild—went on strike last spring, regulating the studios’ AI use was among both organizations’ top priorities.

Yet across sectors, experts we spoke with almost all made the same joke—“AI is not going to take your job, but someone who knows how to use AI probably will.” Which is not all that funny if it’s your job in question. On the bright side, our translation of that quip is: Don’t bury your head in the sand. AI is here to stay, and its capability is pretty awesome, so you would be well served to learn how it can help you. “I think you need to use AI as a tool, but not the end all be all,” says Gina Anderson, the MGS Group real estate agent. In her field, she still sees benefits to “working with an agent who is living the market day to day and sees nuances that AI can’t.”

In some cases, artificial intelligence may lead to job generation in the state. Massachusetts Economic Development Secretary Yvonne Hao says AI has the potential to bolster employment across the key sectors in which we already excel, including life sciences, healthcare, education, and robotics. So perhaps the question isn’t whether the technology will “take your job,” but rather, what opportunities AI can offer you.

Generated by Benjamen Purvis using AI

Is AI Enabling Bad Actors?

You’ve probably heard about some of the scary stuff: AI creating realistic nude images of Taylor Swift or robocalls urging Joe Biden supporters not to vote in the New Hampshire primary. Indeed, “AI has the ability to scale all of the horrible things that we’ve seen carried out on the Internet,” says Laura Manley, executive director of Harvard Kennedy School’s Shorenstein Center on Media, Politics and Public Policy.

That said, Manley doesn’t believe AI—or any technology, for that matter—is inherently good or bad; rather, she describes it as a blank slate that needs regulation and societal guardrails, which are struggling to keep up at the moment, given AI’s lightning-fast development. OpenAI, for instance, recently introduced a highly accurate text-to-video tool that can make any video you ask it to. It’s not hard to see how a technology like that could be used to create disinformation, particularly in an election year.

So how can you determine if you’re looking at a genuine image or something that’s been AI-generated? There are a few ways. Zoom in on the photo or video still to see if you can spot any inconsistencies—are those six fingers on that woman’s hand? Do a reverse Google image search to see if it appears on any legitimate news sites. Run it through an AI image detector tool. And remember: If it looks too good/weird/disturbing to be true, there’s a chance it might be.

Generated by Benjamen Purvis using AI

Will AI Take Over like in a Movie and Control the World?

There is a scene in the film Oppenheimer where General Leslie Groves (played by Matt Damon) asks Robert Oppenheimer (played by Cillian Murphy) the following: “Are you saying that there’s a chance that when we push that button…we destroy the world?” Oppenheimer calmly reassures Groves that the chances are “near zero.” Some of the same fears pervade in the development of AI; in fact, OpenAI CEO Sam Altman has been called the Oppenheimer of our age, a comparison he apparently embraced when he compared his company’s ambition to that of the Manhattan project, according to the New York Times.

The truth is, while the technology is forecasted to change the world in ways we do not yet comprehend, the chances of it taking over the world are extremely low. “Humans are the ones designing and creating AI, so it will always have built-in nuances that will prevent it from taking over in an evil way,” says Giuseppe Stuto, cofounder and managing partner of the tech VC firm 186 Ventures. —Wyndham Lewis

Ask AI: Will AI Take Over like in a Movie and Control the World?

Boston magazine’s human digital editor posed that same question to ChatGPT and asked it to respond in the voice of writer Nora Ephron. Here’s its unedited response:

Well, darling, let’s talk about this whole notion of AI taking over the world, shall we? Picture this: it’s like a scene straight out of a Hollywood blockbuster, complete with robots plotting world domination and humans left scrambling for survival. But let’s not get carried away with the drama, my dear.

Sure, AI has its place in shaping our future, but let’s not paint it as some omnipotent force that’s destined to rule over us like a tyrant. After all, AI is a creation of us mere mortals, and its actions are guided by the intentions we program into it.

Now, don’t misunderstand me; I’m not dismissing the importance of discussing the ethical implications and potential risks of AI. Responsible development is key, darling. We must consider the impact on jobs, privacy, and society at large. But let’s not indulge in the fear-mongering fantasies of AI seizing control and turning us all into mere pawns in its game.

So, my advice? Let’s approach the topic with a healthy dose of skepticism, a sprinkle of wit, and a whole lot of common sense. After all, reality is far more fascinating than any Hollywood script could ever dream up.

Now that’s something, darling. —Camille Dodero

A version of this story was first published in the print edition of the April 2024 issue with the headline, “Boston’s AI Revolution.”