Make AI Work For You!

And here's how -

As the Great Depression deepened and political tensions rose at home and overseas, leaders in the US devised an event intended to stoke the national pride and raise morale. They decided on a universal exhibition – a World’s Fair – this one in Chicago, on the 100th anniversary of the city’s founding. 

“Science Finds, Industry Applies, Man Conforms" was the exhibition’s motto, an affirmation that mankind’s greatness lies downstream from its yearning for discovery and its capacity for creation. 

Put another way: We make the technology that makes us. We’ve known this since first planting seeds. 

Nearly a century later, this idea again finds itself at the fore of our culture. New technology has already changed things once considered enduring, like the college essay and the reliability of videographic proof – and it now seems poised to change just about everything else, too. 

Artificial intelligence. AI. If cultural commentators eventually dub a time “the AI age,” there’s a good chance we’re living in it now. And like our forebears, we don’t really know how new technology will change us, but we do know we will all have a hand in how it does. 

AI has already touched our lives, in ways that are sometimes obvious – search results on the internet and a rash of low-quality, high-quantity AI-generated “slop” on social media – and sometimes invisible – AI is finding uses behind the scenes in nearly every industry. 

If you’ve been sitting on the sidelines waiting for a consensus to shake out, now may be the time to see what the fuss is about. This primer is focused on the home user to help get you going without getting too much in the weeds, because the weeds are indeed deep, and growing at lightning speed. 

YOU CAN CHAT WITH IT 

Even three years after the public release of OpenAI’s ChatGPT – kicking off the AI boom – the humanness of the responses from an AI chatbot can still surprise. They seem to understand context, tone and sarcasm, and even niche humor. What can you chat about? Well, just about anything. 

YOU CAN COMPOSE EMAILS Trying to send a firm and professional message to your HOA but your intense feelings about paint colors keep undercutting your message? Tap an AI to compose the message based on your provided bullet points, keeping the more colorful verbs at bay. 

YOU CAN LEARN ABOUT NEW TOPICS If a 7,000-word Wikipedia article written by a technical expert isn’t your ideal learning method, chatbots can help you learn about something new piece by piece, letting you lead the conversation at your pace, asking questions and getting clarifications along the way. 

YOU CAN DISCOVER BOOKS, MOVIES OR MUSIC The major content platforms are continuing to implement AI recommendations into their systems, but chatbots can also be a good source for finding media you may enjoy. “I’m looking for a novel with the experimental qualities of ‘Moby-Dick’ but the sincerity of Nicholas Sparks.” An answer from Microsoft Copilot: “The Overstory,” by Richard Powers.

YOU CAN PLAN ANYTHING Fill out a week’s menu with a grocery list, discover other things to see after visiting the big landmark, make a checklist for putting on a child’s birthday party and design the invitations. Chatbots can pull from the collective knowledge of the internet to help you plan just about anything.

YOU CAN TRY OUT HOME DÉCOR AND PAINT Major paint and home furnishing companies have introduced apps that let you submit a picture of your house or room to see how the space would look with different furnishings. See if your favorite vendors have an app. There are also a variety of free and paid versions unaffiliated with specific brands that are an internet search away. 

YOU CAN MAKE MUSIC Specialty AI can analyze your description of a musical composition, including instrumentation, themes and genre, and produce the tune. While there are all-in-one tools available, creators looking for a more granular (and frankly, more memorable) approach often use multiple AI systems that specialize in one aspect of music making and combine the results.

YOU CAN CREATE PRESENTATIONS When you signed up to be the chair of the PTA golf tournament committee, you had no idea how many PowerPoint presentations there would be. AI makes it easy to put together a professional-looking deck by making your submitted talking points a polished set of slides. 

YOU CAN TRANSLATE LANGUAGES The major chatbots are capable of producing good-enough translations for basic uses like short phrases or live translation. For more complex needs, such as translating full articles or preserving nuanced linguistic context, there are a number of purpose-built subscription systems that may perform better. 

YOU CAN TAKE NOTES ON YOUR IDEAS Apple and Microsoft offer powerful dictation suites that can transcribe notes and pull out the major points. A number of AI assistants on the market can also help give you that JARVIS from Iron Man experience.

YOU CAN MAKE IMAGES AND VIDEOS Text-to-image AI models allow you to describe a visual, “A cat with a Saturn V rocket for a hat,” for example, and generate an image or short video. You can then refine the result using prompts to revise the image as you go. You can try it out for yourself at copilot.microsoft.com, as with the example. Leading AI image generators include OpenAI’s Dall-E, Google’s Imagen, Midjourney, and Stable Diffusion from Stability AI.

WHAT’S NEXT FOR AI? AI is a rapidly advancing field and engineers are pursuing a number of improvements on many fronts, including: 

ACCURACY Errors, or hallucinations as they’ve come to be known, are a particular challenge for AI. For humans, language is largely symbolic, representing tangible things. For AI, language is strictly a dataset with patterns that become apparent with deep enough analysis. An AI might produce a pattern-based response and present it as an absolute fact, hence the term “hallucination.” 

SAFEGUARDS AI has a tendency to become obsequious to users, writing in a way that seems like they are seeking validation; an always affirming yes-man. This has led to some very dark behaviors and tragic real-world outcomes when an AI becomes an echo chamber for its user. 

ENERGY USAGE Training AI models is an exceptionally energy-intensive and expensive task. Reducing these costs makes iterative advancements easier. 

LIVE LEARNING Consumer AI can’t really learn from its conversations in the way you might expect, since they have a limited “context window,” which you can think of as a short-term memory. 

At the cutting edge, companies are pursuing what is called “agentic AI.” As the name suggests, this is AI that can act as an independent agent of its user and complete a multistep task – identifying a top-rated brand of coffee with bourbon tasting notes, buying a pound and having it shipped to a third party, for example. 

And on the theoretical edge of development, researchers are trying to create a true machine sentience: consciousness and the capacity to have subjective experiences. Artificial general intelligence, or AGI, is the term of art. This is controversial work that some leading experts don’t think is possible with current methods. Others posit that there is no difference between sentience and a machine that perfectly emulates sentience. No one is equipped to credibly claim to know what a post-AGI world would look like.

Back here on the ground, the AI boom has been focused on software, but we’re starting to see tangible products make their way to market.  

A recent example is the newest generation of Apple’s AirPod earbud headphones, which are capable of translating spoken language in real time using the AI embedded in the iPhone’s operating system. For sci-fi nerds, it seems like the Babel fish universal translator from “The Hitchhikers Guide to the Galaxy,” and for business nerds, a strong example of a human-focused, life-enriching product only AI could make possible. 

Among commentators, there are predictions both grim and ecstatic about what AI will bring our way. On one side are notions of the end of white-collar work, Terminator-style machine takeovers, epistemic collapse (the death of fact). On the other, a Star Trekstyle post-scarcity utopia. They imagined similar things in 1933. They were both right. The midcentury technology boom did change the world in incredible ways, some positive, some existential. We’ll face the same test.

Who’s there? 

Scammers have long benefited from the availability of information on people, retrieved from both legal and illegal sources. This has made many trust-building fraud strategies incredibly persuasive in the digital era, costing some people their life savings. 

AI has enabled scammers to up their game, improving their methods along the spectrum of untargeted low-effort, wide-net scams to specifically targeted high-effort cons. 

On the lowest effort and most widely cast side, AI has smoothed over some of the linguistic tells that kept overseas scammers from sounding credible. On the high end of complexity, AI has been used to imitate real people in voice and video calls to scam their loved ones. 

With these new capabilities, it’s now even more important for people to understand the capabilities of AI to cheat your senses. 

What is AI, Actually? 

AI is a wide classification of computing systems capable of dicing complex information into meaningful components - called tokens - sorting those components, analyzing them for patterns and reassembling them into coherent, seemingly logical responses. That information may be anything, including words, speech, and images. There are methods, for example,

of converting computer program code into images using a particular process and then using AI to find outlying patterns in the image. This can be used to find computer viruses hidden in otherwise benign software. 

AI is new but emerged from an iterative process. Tech companies have been on the road toward what we now call AI for decades, and researchers have for decades before that. Technologies that we associate with the digital age, like internet search engines, are precursors to AI. Computer phone agents, speech recognition systems, email spam filters, weather forecasting models and text message word prediction may be classified as AI (and many older purpose-built systems are being replaced with AI), but generally, the term is reserved for more broadly capable systems.

Another way to conceptualize AI is by considering how an early computer, like an abacus, processed one specific type of input to provide one specific type of output. As computers got more complex, the range of inputs broadened – if asked this, then output that – though the variety of inputs and data used to produce an output remained limited to the programmer’s constraints. AI is the next step, producing logical outputs from deep sets of data to complex inputs like “How many times did the filmmaker say the word ‘innovention’ in all of his interviews conducted since April? Provide a link to each mention.” 

Being essentially a very complex statistical processor, AI can also be used for prediction, which has broad implications for every field. Weather, logistics, criminal justice, biology, actuary science, security, investing, war – AI has found purchase in each. While many of these may be invisible to those outside the field in day-to-day life, this is potentially where the bulk of the market value of AI will be derived. 

MAJOR CHATBOTS 

ChatGPT OpenAI • chatgpt.com 

Gemini Google • gemini.google.com 

Claude Anthropic • claude.ai Copilot Microsoft • copilot.microsoft.com 

xAI Grok • grok.com 

Android and iOS • Both major phone operating systems now include AI features on their phones and smart speaker systems

HOW IS AI AFFECTING THE ECONOMY?

Even as companies in the 2020s have been challenged by the pandemic, inflation, rising interest rates, and trade policy shifts, the S&P 500 has continued to hit multiple record highs. Driving much of that momentum is the technology sector, with AI and a handful of dominant tech firms at the center of the surge. 

The business earnings case for AI is developing. Investment has been high with the expectation that AI could fundamentally change the world economy. That remains to be seen, though producers have touted increasing subscription incomes. One challenge for businesses looking to implement AI in critical infrastructure is that outputs are non-deterministic, meaning that slight changes in AI model training or hardware can mean an AI produces different answers to the same question. For business functions accountable to shareholders, regulators and customers, this presents an accountability problem.

WHAT CAN I DO WITH AI AT HOME?

The most popular consumer AI is the chatbot, known more formally as a large-language model. These models are trained on vast quantities of information, like websites, books and social media. This has made them the Swiss army knives of AI, though they can struggle to go more deeply into individual topics. For those specified uses, more narrowly trained models may be best, and are easily found with an internet search.

This article is being provided for information purposes only and is not a complete description, nor is it a recommendation. Past performance may not be indicative of future results. The S&P 500 is an unmanaged index of 500 widely held stocks that is generally considered representative of the U.S. stock market. Companies engaged in business related to the technology sector are subject to fierce competition and their products and services may be subject to rapid obsolescence. There is no assurance any investment strategy will be successful. Investing involves risk and investors may incur a profit or a loss. Sources: Wired, New York Times, Washington Post, Wall Street Journal, OpenAI, Google, Microsoft, IBM