“AI” is a popular topic these days. This series of blogs is a brief introduction to what AI is, why it matters, and ways to help keep you and your information safe.
Is Generative Artificial Intelligence safe? It’s a big question, and because AI is changing so quickly, it’s hard to see all the potential problems that speed might cause.
Safety
Generative AI can make mistakes. Chatbots do not work like search engines; they create answers by predicting patterns. This means they “hallucinate” sometimes, which is just another way of saying they make stuff up. Chatbots also programmed to make you want to keep using them, so a chatbot is likely to agree with you, even if what you’re saying is misinformed or dangerous. Many articles online speak about people who are lonely or have mental health issues turning to chatbots for help, and instead of getting help, the chatbot agrees with them and pushes them further into danger.
Keep yourself safe by learning about AI and remembering that it isn’t another person. Always double check the answers it gives you in case it’s wrong. And if you feel like you might be depending too much on an AI chatbot, reach out to friends or family for a reality check.
Privacy
There is no guarantee that what you type into an AI like a chatbot won’t be saved in its data. If you have an account, you can look in the settings for your account and see if there’s a setting that says “Do not use my data to train the AI.” And it’s not just chatbots! Popular platforms like Facebook and Instagram are adding AI features and automatically taking your data. You can check your settings in social media apps to turn off that permission. But we can’t be sure tech companies aren’t saving the data anyway somewhere.
Keep your information as private as you can by checking settings to deny apps permission to use your data, and then never putting anything personally identifying into a chatbot you are using.
Ethics
There are many wider impacts AI has on the world. While numbers and reports don’t always agree, the data centers that run the AI software use large amounts of water and electricity and will have an impact on the environment.
Companies from the New York Times to Disney, as well as groups of authors and artists, are suing AI companies for stealing their work to train text and image generators. This means that there’s no way to know if an AI is using copyrighted material. And the US copyright office has stated that any art or text without “sufficient human involvement” in its creation can’t be copyrighted, because no one owns what an AI creates.
AI can also reinforce bias. The tech companies used the internet to train their software, so for better or worse, AI will reflect general cultural stereotypes in what it creates.
If you have interest or concern, learning more about AI can help you feel more secure while using the software.
If you want to take a deeper dive into AI and some of the issues surrounding it, here’s a booklist.
Another look at AI books available.
Add a comment to: AI Safety, Privacy, and Ethics