Artificial intelligence is a phrase that can inspire awe or fear, depending on who you talk to.
But whether we like it or not, it looks like AI is here to stay. According to a forecast by MarketsandMarkets, the total market is expected to reach $1,339 billion by 2030. A Forbes survey found that 77% of respondents were concerned that AI would cause job losses.
This fear is valid, says Manasi Vartak, chief AI architect at Cloudera, but it can ultimately hold employees back. Instead, you risk losing your job to someone who is more open to embracing new technology.
What is generative AI?
Generative AI helps us “go from imagination to reality,” says Joe Edwards, director of product marketing at automation software company UiPath.
Generative AI can create words, music, images or videos with just a few suggestions. It has caused a stir on social media this year as AI art, fake celebrity images and posthumous music began circulating.
Generative AI systems are trained on large amounts of data, studying photos, videos and the way people operate online, Edwards says. These advanced learning machine models can then identify patterns and create their own content.
“AI used to be what we call ‘predictive’, it can tell whether a tweet or a news item is positive or negative,” says Vartak. “But now you can write an entire news story, creating content from scratch, and that’s why this is so different and amazing.”
Generative AI has both strengths and weaknesses.
For example, he is great at writing, Vartak says. It can be a tweet, an email, or creating an elaborate, fantastic story. Sometimes it can break down complex topics to help you summarize or learn information. It can also perform simple everyday tasks such as transcribing meetings and sorting data.
But it can often become too fanciful and give you incorrect or misleading information. For example, if you ask ChatGPT, “How many r’s are in the word strawberry?” you will repeatedly receive wrong answers.
“It can sound very confident when it’s lying to you,” says Edwards.
When a large language model perceives non-existent patterns or spits out nonsensical answers, this is called “hallucinating.” It’s a big challenge in any technology, Vartak says. If you ask him to summarize an article or paper, he may only be 80% right.
Some wrong answers are low stakes and quite funny, while others can spread dangerous misinformation. When Google debuted its AI overview earlier this year, one viral response read: “Doctors recommend smoking 2-3 cigarettes per day during pregnancy.”
And because it is a man-made creation, generative AI can amplify human biases. AI images can perpetuate harmful racial and gender stereotypes, the Washington Post found in 2023.
“These models are trained on data that exists in the world and is biased, so they may not include representation of women, of minorities, of LGBTQ (communities), of people of color,” Vartak says. “If we just take what technology tells us as gospel, we will miss those stories completely.”
How to use generative AI
Keep a human informed, Vartak says. Generative AI should be your navigation partner, not your driver.
“Give it some ideas, let it generate some text, then go look at it and make sure it makes sense,” she says. “Trust but verify.”
Edwards used a generative AI program to sort through the pile of emails he had after returning from paternity leave. You can also use it as a brainstorming partner when planning your child’s birthday party or an upcoming trip.
Aspiring chefs can try it out in the kitchen to help generate recipes. Musicians can experiment with custom track makers, and some potential homeowners use it to buy a house.
There are also business applications that can be tailored specifically to your work. Unfortunately, for teachers, this means developing a keen eye for ChatGPT-generated student essays and responses. In healthcare, some doctors are using AI to improve patient visits and translate visit notes into understandable terms. Some AI systems can detect breast cancer, pulmonary embolism and stroke and can lead to life-saving earlier diagnoses.
Software companies like Cloudera and UiPath are creating custom private AI systems trained on smaller amounts of data to prevent leaks and hallucinations.
Both Vartak and Edwards recommend trying out generative AI – it can be as low-commitment as playing with ChatGPT or a more involved free online training course.
“Think of it as your sidekick,” says Vartak. “You will be able to use it to… have more fun and be more productive in your daily life. I think all new technology is difficult, but this one is here to stay and can have a positive impact.”
AI can spread disinformation: Here’s how to spot it online
Just curious about more? We’ve got you covered.
USA TODAY explores the questions you and others ask every day. From “How to Lucid Dream?” to “What did Albert Einstein invent?” to “What is the most spoken language in the world?” – we aim to find answers to the most common questions you ask every day. Head over to our Just Curious section to see what else we can answer for you.
