Artificial intelligence has entered a new era. From search to education to the arts, recent advances in AI promise to revolutionize the way we work and live.
But for some of us, AI raises more questions than it answers. How does it affect us? Are there any risks? How can we make it reliable?
Before answering these complex questions, it helps to understand the basics of AI. Read our crash course on artificial intelligence.
First of all, what exactly is AI?
AI is essentially software that can learn patterns from information. Think about things like language, images, sounds, and online behavior. AI uses patterns in existing and new data to perform tasks that typically require human intelligence, such as finding products you might buy or completing the sentences in an email. Make predictions to
How does it work?
For example, consider a customer service chatbot. When you enter your question, the chatbot uses an algorithm (or set of rules) to recognize keywords and determine the type of help you need. Machine learning models generate appropriate responses based on existing and new information. Chatbots improve over time as they interact with new customers and receive more data.
As Mozilla researcher Becca Ricks explained, “Think of an algorithm as a program that manipulates a dataset. A model is the output that makes a prediction.”
Why do chatbots like ChatGPT sound more human-like?
Modern chatbots use a type of machine learning model called a neural network. Inspired by the structure of the human brain, it is designed to learn increasingly complex patterns to derive predictions and recommendations. Chatbots allow models to learn language from large amounts of existing and new data and are very good at hearing how people speak.
Can AI get things wrong?
absolutely. AI models learn from data, which can be incomplete. For example, ChatGPT is a language model trained on data from the Internet. Therefore, solving simple math problems can be difficult.
AI can also produce biased output. For example, image recognition trained on a set of images that primarily features light-skinned people may not be able to recognize darker-skinned individuals. Because the algorithms and data are provided by humans, AI technologies are typically subject to existing biases, such as race-, gender-, and age-based biases.
“They can affect whether our friends see our posts,” Becca says. “Or it may affect whether or not you can get resources from your local government.”
How can we know if we can trust AI?
The first step is to learn about it. From there, you can demand transparency and accountability.
“The more we all know about how AI systems work, the easier it will be to imagine something better,” Becca says. “And it becomes easier to design alternatives that benefit society and reflect the values of the community.”
If you want to learn more about AI, the people who create it, and how it's impacting our communities, check out Mozilla's latest season. IRL podcast.
AI is present in almost everything you use on the internet, including search engines and social media feeds. However, if you want to reduce the amount of personal information exposed on the web, avoid revealing your real email address and phone number when signing up for modern AI apps.
Like any other application, AI apps can expose information to online trackers, spammers, and hackers. Firefox Relay provides email and phone number masking, so you can sign up for a new account anonymously.
Start protecting your email inbox today
Sign up for Firefox Relay
