Machines Get Smarter

Machines Get Smarter

by Rick Coleman

a self-driving vehicle

Artificial intelligence is improving. Some wonder if

that is good news or bad news.
Movies-from I, Robot to Avengers: Age of Ultron-have tried to warn us: as soon as computers can
think for themselves, they will decide that the best thing they can do for the world is to kill human
beings. Of course, that hasn't happened-but computers are becoming more and more intelligent.

Is that a problem? Is it time for humans to worry?

How Intelligent Are Intelligent Machines?

The science of getting machines to think like humans is called artificial intelligence, abbreviated as AI.
The most common forms of AI are complex mathematical formulas known as algorithms. For
example, the program that answers the phone when you call a company may sound like a human
being, but it's merely following a set of rules in an algorithm. Likewise, when your music-streaming
service makes a recommendation, it simply has had its algorithm compare your previous "likes" to
those of millions of other users.

These are examples of "machine intelligence." Machines-computers-today have the ability to

quickly digest vast amounts of data and sort the results according to complicated rules. But that
doesn't make them intelligent in the way that humans are.
ReadWorks.org · © 2020 ReadWorks®, Inc. All rights reserved.
 Machines Get Smarter

In fact, the most intelligent thing about the AI program that answers your calls is actually its ability to
understand you. In order to work effectively, speech-recognition programs must make allowances for
the wide variety in human pronunciation and vocabulary.

Recognizing and identifying individual people by sight is far more difficult. While humans can easily
tell one person from another, computers are only beginning to have that ability. However, machines
that can identify people are fast becoming a reality: facial-recognition software is already in use by
many police departments around the world.

Another major development is the arrival of self-driving cars-vehicles that use even more complex
algorithms to interpret video signals and identify roadways, obstacles, other cars, and pedestrians.
Cars governed by these programs really do act independently.

Learning to Learn
Advanced AI takes things even further. Today, programs can "teach" themselves-learning from their
mistakes, changing the way they process data to improve their performance. The programs can then
act without any further input from human beings, making decisions based on data. This process is
called "machine learning."

The question that emerges from these developments is a simple one: can computers truly think for
themselves? We cannot say yes today, but there are signs that point in the direction of "Yes,
eventually." For example, in March 2020, scientists used AI to analyze 40,000 articles about the
coronavirus in order to find trends in research that might lead to a treatment. That's pretty close to
coming up with an original idea-and what's truly remarkable is that the algorithms used were not
written specifically for that purpose.

Those same programs can be used to study any large set of documents, on any subject. If those
programs were set loose, what would they study and what ideas would they come up with? And if
they were put in charge of our self-driving cars, where would they take us? What songs would they
play on the sound system? It seems as if we are going to find out sooner rather than later.

We can only hope that no one lets them watch Avengers: Age of Ultron.

ReadWorks.org · © 2020 ReadWorks®, Inc. All rights reserved.

 ReadWorks Vocabulary

Vocabulary

analyze
verb
definition: to separate into parts for close study; examine and explain.
If we analyze the problem, perhaps we can solve it.

Spanish: analizar
forms: analyzed, analyzes, analyzing

artificial
adjective
definition: made by human beings; not natural.
He can walk now with an artificial leg.

Spanish: artificial, postizo

interpret
verb
definition: to decide on or explain the meaning of.
How do you interpret his latest book?

Spanish: interpretar
forms: interpreted, interpreting, interprets

ReadWorks.org · © 2023 ReadWorks®, Inc. All rights reserved.

Definitions and sample sentences within definitions are provided by Wordsmyth. © 2023 Wordsmyth. All rights reserved.
 Name: __________________________________ Date: _______________

1. Before you start reading...

Here are the vocabulary words that will be in this reading. Let's see how well you already know them.

Check the box that shows how well you know each word. It's ok if you don't know them yet (this is not
graded)!

Have heard of it but

Know something
Don't know it not sure of its Know it well
about its meaning
meaning
analyze
artificial
interpret

2. Word Matcher
Every word has other words that have similar meanings or even the exact same meaning (these are
called synonyms!). Draw a line from each similarword or synonym to the vocabulary word that it
matches!

break down dissect clarify man-made synthetic

analyze artificial interpret

manufactured clarify explain comprehend

3. After reading and exploring the words through some activities...

Do you know these words better? Check the box that shows how well you know each word. It's ok if
you don't know them yet (this is not graded)!

Have heard of it but

Know something
Don't know it not sure of its Know it well
about its meaning
meaning
analyze
artificial
interpret

ReadWorks.org · © 2023 ReadWorks®, Inc. All rights reserved.

Vocabulary powered by WordSmyth. © 2023 Wordsmyth. All rights reserved.
 Machines Get Smarter - Comprehension Questions

Name: ___________________________________ Date: _______________

1. What does AI stand for?

A. artificial indoctrination
B. altered intelligence
C. artificial information
D. artificial intelligence

2. How does the author describe "machine intelligence?"

A. as a computer's ability to come up with original ideas and thoughts, much like an
intelligent human would
B. as a computer's ability to digest massive amounts of data quickly and sort the results
according to complicated rules
C. as a computer's desire to learn new facts and figures because it can think
independently and loves knowledge
D. as a computer's personality that develops over time as it learns more facts and solves
more equations

3. Read the following sentences from the passage.

"Advanced AI takes things even further. Today, programs can "teach" themselves-
learning from their mistakes, changing the way they process data to improve their
performance. The programs can then act without any further input from human
beings, making decisions based on data. This process is called 'machine learning.'"

What can you conclude about machine learning based on this information?

A. Programs that use machine learning probably perform better than programs that
depend on humans for corrections or improvement.
B. Programs that use machine learning work the same as programs that depend on
humans for corrections or improvement.
C. Programs that use machine intelligence are less reliable for data processing than
programs that rely on humans for corrections or improvement.
D. Programs that use machine learning are less popular with data scientists than
programs that rely on humans for corrections or improvement.

ReadWorks.org · © 2023 ReadWorks®, Inc. All rights reserved.

 Machines Get Smarter - Comprehension Questions

4. How do you think the author would describe the future of AI?

A. as likely to stay the same

B. as likely to keep getting smarter
C. as probably not relevant to the future of humans
D. as probably bad for the future of humans

5. What is the main idea of this passage?

A. At first, machines had a difficult time recognizing different human faces, but they have
recently become much better in facial recognition technology.
B. Algorithms have already helped machines perform complex tasks, and as machine
learning gets more advanced, machines are getting closer to being able to think, too.
C. The algorithms behind the systems that answer companies' phones have to be able to
interpret a wide range of pronunciations and vocabularies in order to work well.
D. In March 2020, at the start of the coronavirus pandemic, researchers fed AI 40,000
articles on research trends to help them search for a possible treatment.

6. Read the following sentences from the passage.

"For example, in March 2020, scientists used AI to analyze 40,000 articles about the
coronavirus in order to find trends in research that might lead to a treatment. That's
pretty close to coming up with an original idea-and what's truly remarkable is that
the algorithms used were not written specifically for that purpose."

As used in the sentence, what does the word "analyze" most closely mean?

A. examine and interpret something

B. ignore and hide something
C. examine and destroy something
D. remember and make note of something

ReadWorks.org · © 2023 ReadWorks®, Inc. All rights reserved.

 Machines Get Smarter - Comprehension Questions

7. Please choose the answer that best completes the sentence below.

Scientists used AI to analyze coronavirus research trends, _____ the algorithms used
were not built for that specific purpose.

A. in conclusion
B. for example
C. even though
D. eventually

8. Describe the process of machine learning. Use details from the text to support your
description.

9. Which involves more human input, basic algorithms or machine learning? Support
your answer with evidence from the text.

ReadWorks.org · © 2023 ReadWorks®, Inc. All rights reserved.

 Machines Get Smarter - Comprehension Questions

10. Imagine you were building a robot that could perform surgery on its own. Would you
rather have it rely on a basic algorithm with a human supervisor, or on a program
that uses machine learning? Support your argument with details from the text.

ReadWorks.org · © 2023 ReadWorks®, Inc. All rights reserved.