Artificial intelligence has, for decades, been fodder for sci-fi movies, philosophers, and sleep-deprived computer programmers, but suddenly it seems to be everywhere.
ChatGPT has reached 100 million users at an unprecedented rate. Bill Gates recently declared that “the era of artificial intelligence has begun.” Last month, the Biden administration began exploring new measures to hold AI systems accountable for their impact.
But for many people, the concept is still a vague one that does not affect their daily lives.
So this might be a good moment to step back and review the basics. Here’s a guide to help you understand what all the hype is about.
Why is everyone suddenly talking about AI?
You can thank (or blame) one particular company: OpenAI, a San Francisco-based tech startup with a few hundred employees. In November, OpenAI released ChatGPT chatbot software to the public, and it quickly became apparent that it was far ahead of the chatbots that had come before. It was like talking to someone who knows everything.
The tool, which the company says is just one step in a long process of developing artificial intelligence, quickly went viral. Other tech companies, such as Google and Meta, have been testing similar chatbots behind closed doors, but OpenAI has made them widely available — a controversial decision due to unknown risks.
What’s so great about a chatbot?
Medium chatbots have been around for a long time. Think of the customer service chat windows that appear on some websites. In 2016, Microsoft released an AI chatbot named Tay, but quickly scrapped it after people taught it to use racist language.
ChatGPT came on the scene as something different. Not only can she answer a seemingly unlimited number of questions, but she can also write scenarios, summarize vast amounts of information, and imitate a human in conversation in a rather convincing manner. It seemed right away that, at least, it could one day make everyday life more efficient.
And chatbots are just one part of AI, along with animated images and videos, facial recognition technology, and more.
Let’s go back. What is artificial intelligence even?
At its simplest, AI can be summed up in a few words: machines that think. Or, even better, machines that can tradition thinking.
The term has its origins in scholars after World War II. British mathematician Alan Turing in 1950 predicted the development of “digital computers” that could convincingly imitate humans, and in 1955, American mathematician John McCarthy and colleagues at Dartmouth College coined the term “artificial intelligence” in a research proposal.
Generative AI, a new term, refers to software like ChatGPT that generates new material. You can find a more comprehensive glossary of AI terms here.
Is it really possible for computers to “think”?
We could write an entire book about this one, but here’s a short answer: No, they can’t. While few people think that artificial intelligence has really taken off, they are a small group, and the idea is really to deflect attention from what’s going on inside computers.
If you’d like a longer answer, NBC News spoke with several philosophers about how they approach the question.
The AI program is able to convincingly imitate humans because it is good at prediction: it guesses which word, sentence, or image you want to see next. (Some critics have called this “Glorious autocomplete. “)
And the systems are so good at predicting because their human creators fed them many previous human-created examples—including huge chunks of the internet. The raw material that goes into AI models is called training data, and although some companies are secretive about what they use, known data sources include Reddit and Wikipedia.
Okay, so AI is drawing insights from a lot of data. how?
Artificial intelligence learns by example. Looking at us, language models identify patterns in how we write and speak, distilling concepts like tone, word placement, and even idioms. These patterns are then translated into mathematics in a process called “modal training”. Like children learning new words and grammar, AI must understand the rules of engagement.
When large language models like ChatGPT receive prompts, this knowledge allows them to understand what we are asking and build responses.
ChatGPT takes training even further with its secret link: reinforcement learning from human feedback, or RLHF. Fine tuning technology does the heavy lifting. At this point, human classifiers score the model’s output, severely penalizing wild, inappropriate, or completely meaningless answers, while rewarding those that are useful and human-like. This enables smooth conversation exchanges.
While there are other tuning techniques, RLHF has been considered a pioneer in language modeling, and is used by companies such as OpenAI or Hugging Face, a startup that provides tools for programmers to build their own AI models.
Is artificial intelligence just another Silicon Valley fad?
The tech industry has been rocking one fad after another lately, from self-driving cars and metaverses to NFTs and web3.
On one level, AI chatbots may bear some resemblance to those disappointing ideas — do we all want to spend our days talking to a computer? But there are reasons to believe that AI is more than just another passing trend.
For one thing, money is pouring into the sector, with $1.7 billion in startup investments alone in the first three months of 2023, according to research firm PitchBook. Plus, concrete uses have already emerged, from hit songs to helping the blind.
Why is all this happening now?
It has been 26 years since IBM’s Deep Blue computer program defeated chess champion Garry Kasparov – a milestone in artificial intelligence research and development. Since then, computer chips have become much faster and can handle the huge amount of data required for modern artificial intelligence, and new ways of writing programs have made the process more efficient.
Chipmakers like Nvidia and technology companies including Google, Meta, and OpenAI have poured resources into these two areas, as well as into integrating talented computer scientists under their respective roofs.
When can I expect this to start affecting my life?
Don’t expect to wake up one morning and suddenly live in the world of artificial intelligence. Instead, expect changes to come a little at a time: a hit song created with AI, a new test in the doctor’s office to detect cancer or a little better customer service. OpenAI has licensed its technology to Morgan Stanley so its investment advisors can offer better advice and Khan Academy so its students can access a chatbot tutor.
Think of all the businesses or products you deal with every day, and there’s a good chance that one is using similar technology or will be using it in the near future – even if the only immediate effect is a slightly increased efficiency.
Can we expect any major changes?
It’s hard to tell what to count on, but yes, there are a lot of dreams in AI startups. If AI software can make both human and computer work more efficient, can all that brainpower be put into making breakthroughs in other new areas?
There are two areas where there is a lot of optimism: Pharmacy shelves are full of new drugs designed for AI and AI software that can enable new power plants based on cleaner fusion energy.
Will AI make a lot of jobs irrelevant?
Predictions run the gamut, so if you’re confused, you’re not alone. Sam Altman, CEO of OpenAI, has suggested that AI will lead to a utopia in which people don’t need to work, while others warn of mass unemployment among computer programmers.
Even labor economists are baffled, advising that AI will change people’s jobs and complement existing work but otherwise avoid specific predictions.
A group of researchers recently attempted to rank jobs according to the risk that AI will change what people do. In trouble, they say: telemarketers, humanities professors, and credit-issuers. Hard to replace: the dancers, the builders, the steelworkers.
And who will make money from this?
Again, expectations are all over the place, from a more equitable society to a less equal one. Much depends on the reaction of politicians and voters, and the Biden administration and Congress are paying increased attention to AI research and development.
But some of the early leaders are big tech companies, such as Google, Meta, and Amazon. OpenAI, which in 2019 transformed from a non-profit organization into a for-profit company; And whoever survives is among the dozens of AI startups that collectively raise billions of dollars from early adopters.
What could go wrong?
If you go by sci-fi movies or the nightmares of a few researchers, there’s a chance killer robots: AIs become sentient beings with motivations of their own.
Based on this scenario, thousands of people, including Elon Musk and some AI researchers, have signed a petition calling for at least a six-month pause in training new AI systems. But some senior technology executives and researchers did not sign on. And at least so far, there is not much information indicating that humans are in direct danger due to artificial intelligence.
So how anxious should I be?
It depends who you ask.
Most of the immediate risks relate to short-term abuse by humans, not bots. There is ongoing research on using AI to hack people’s passwords, and the Washington Post has revealed someone using an AI-generated image as a thirst trap, possibly in exchange for money.
One thing to watch: how quickly we see progress in physical robotics. Hardware hasn’t evolved as much as software, and a couple of years ago OpenAI disbanded its robotics team even after it got a robot to solve a Rubik’s Cube. But OpenAI is now investing in a Norwegian robotics company.
#understand #artificial #intelligence #guide