5.1 Beneficial and Harmful Effects

A computing innovation is anything new that includes a program as a key part of how it works. It can be a physical device (a smartwatch, a self-driving car), a nonphysical piece of software (a photo editor, a messaging app) or a concept (e-commerce, online banking).

Innovations are made by people, and once they exist, people change their habits around them. Maps apps replaced paper maps and printed directions. Online calendars replaced wall calendars for many families. Streaming changed how people find and pay for music.

Computing has also opened new ground in other fields. Doctors use programs to spot patterns in scans, engineers simulate bridges before building them, and artists make digital music, animation and interactive installations that weren't possible before.

An effect isn't automatically good or bad. Whether it helps or hurts depends on who you ask, and sometimes the same person feels both ways.

Take a fitness tracker that shares your step count with friends. You might love the friendly competition and the motivation to walk more. The same feature can feel like pressure, and it lets other people see when you're active and when you're home.

When you explain an effect, say who is affected and how. "Social media is bad" earns nothing. "A video app's recommendation feed keeps teens watching longer, which helps the company sell more ads but can cut into the students' sleep" names the effect, the people and the result.

Creators can't foresee every use of what they build. Some unplanned effects are welcome; others cause real harm.

The World Wide Web began at CERN, the European physics lab, as a way for researchers to share documents quickly. Nobody designing it was planning for online shopping, video streaming or social networks, yet all of those grew out of it.

Targeted advertising helps a small business reach people likely to want its product. But the same data that powers it can be misused, for example to push misleading political ads to a narrow group or to show certain job or housing ads only to some kinds of people.

Machine learning and data mining help find new drug candidates and catch credit card fraud. The same pattern-finding methods have also been used to treat groups of people unfairly, such as charging some neighborhoods more for insurance.

Because a program can be copied and shared instantly, its effects can grow far beyond what one developer controls. An app that reaches millions of people in a week spreads its benefits fast, but also its bugs, its privacy problems and any misuse.

Responsible programmers try to think through how their work might be used, including uses they didn't intend, and to plan for the harmful ones. They still can't catch everything, which is why testing with real users and listening to feedback after release matter.