This article is part of the On Tech newsletter. You can register here to receive it weekdays.
When you hear about artificial intelligence, stop imagining that computers can do everything we can but better.
My colleague Cade Metz, who has a new book on AI, wants us to understand that the technology is promising but has its downside: It is currently potentially inferior to humans and it is being encrypted. with human prejudice.
I spoke to Cade about what artificial intelligence is (and not what it is), the areas where he hopes and fears the consequences, and the areas where AI doesn’t have optimists’ hopes .
Shira: Let’s start with the basics: What is AI?
Cade: It’s a term for a set of concepts that allow computer systems to function vaguely like a brain. Some of my reports and my book focus on one of those concepts: a neural network, which is a mathematical system that can analyze data and identify patterns.
For example, if you take thousands of pictures of cats and put them in a neural network, it can learn to recognize patterns that define the shape of a cat. The first neural networks were built in the 1950s, but for decades they never really fulfilled their promises. That started to change around 2010.
What has changed?
For decades, neural networks had two significant limitations: insufficient data and insufficient computer processing power. The Internet provides us with tons of data, and in the end scientists have enough computational power to handle it all.
Where can anyone see the effects of neural networks?
This one idea has changed many technologies over the past 10 years. Digital assistants like Alexa, unmanned cars, chat bots, poetic computer systems, surveillance systems, and robots that can pick up products in a warehouse rely on neural networks.
Sometimes it feels like people talk about artificial intelligence as if it’s a magic potion.
It’s correct. The original sin of AI pioneers was that they called it artificial intelligence. When we hear this term, we envision a computer that can do anything that humans can do. That was not the case in the 1950s, and it is not true now.
People fail to realize that it is difficult to copy our human reason and our ability to deal with uncertainty. Self-driving cars can perceive what’s around them – in some ways better than humans can. But it doesn’t work well enough to drive anywhere anytime or do what you and I do, like reacting to something surprising on the road.
Disadvantages from neural networks and AI?
So many, so much. Machines will be able to generate misinformation on a large scale. There won’t be any way to distinguish what is real online and what is fake. Self-propelled weapons are also incredibly dangerous.
And the scariest thing is that many companies have pitched algorithms as the utopia that can eliminate all human errors. It’s not. Some neural networks learn from vast amounts of information on the internet – and that information is produced by humans. That means we’re building computer systems that show human bias – against women and people of color, for example.
Several American technologists, including former Google chief executive Eric Schmidt, To speak that the United States does not take AI seriously enough and that we risk falling behind China. Really spoiled that concern like?
It’s legal but complex. Schmidt and others want to try to ensure that the most important AI technology is built inside the Pentagon, not just inside tech giants like Google.
But we have to be careful about how we compete with a country like China. In the United States, our best tech talents often come from overseas, including in China. Closing our borders with experts in the field will hurt us in the long run.
Tips of the week
How to become a savvy online shopper
A reader named Eva emailed On Tech asking about small software programs called browser extensions, plug-ins or add-ons for Chrome, Safari, and Firefox claiming they would save money. she.
“I continue to see ads for these browser add-ons like Honey (from PayPal) and Capital One Shopping,” she wrote. “They claim that they will automatically find and apply promo codes to save you money whenever you shop online. This sounds great, but I keep wondering, What does it have for them? They don’t just do this for their kindness. Before I sign up for these services, I want to know what the tradeoffs are. Can you help me find out? “
Brian X. ChenThe New York Times personal technology columnist responded as follows:
Yes, there is always a trade-off. With free software, your personal data is often part of the transaction.
Read between lines: That means Honey can track your browsing on retail sites. (Honey has said that it only uses data in ways people expect.)
That is a lot of information to hand over to the coupon software automatically applied. Is it a fair trade or not is up to you.
Before we go …
So. A lot of. Money. Everywhere: My colleague, Erin Griffith, connects dots between digital art that sell for $ 69 million, a crypto frenzy, and the skyrocketing price of things like classic sneakers. . Basically, taking financial risk right now comes at a cost, plus our brains are getting worse in a pandemic. Related: Stripe, the company that creates software plumbing for businesses that accept digital payments, is now one of the most valuable startups in history.
Facebook is working on our vaccine views: Facebook is conducting internal research on spreading ideas across its applications that contribute to vaccine hesitation, The Washington Post reported. Early findings suggest that messages that are not completely false can “harm certain communities where it works to suppress echoes,” The Post said.
How to keep Americans safe: The failure of the US intelligence services to detect the recent digital attacks by Russia and China is causing US officials to rethink how the country should protect itself, its colleagues. I report. One tricky idea is for tech companies and US intelligence agencies to collaborate on real-time assessments of cyber threats.
Go hugging a cow. It will help.
We want to hear from you. Let us know what you think about this newsletter and what you want us to explore. You can contact us at [email protected]
If you have not received this newsletter in your inbox, Please register here.