Sekėjai

Ieškoti šiame dienoraštyje

2024 m. spalio 9 d., trečiadienis

How to Say No to Our A.I. Overlords


"Google, Microsoft and Meta are shoving A.I. chatbots into our faces. Sometimes, there’s a way out.

Big tech brands like Google, Apple, Microsoft and Meta have all unleashed tech that they describe as artificial intelligence. Soon, the companies say, we’ll all be using A.I. to write emails, generate images and summarize articles.

But who asked for any of this in the first place?

Judging from the feedback I get from readers of this column, lots of people outside the tech industry remain uninterested in A.I. — and are increasingly frustrated with how difficult it has become to ignore. The companies rely on user activity to train and improve their A.I. systems, so they are testing this tech inside products we use every day.

Typing a question such as “Is Jay-Z left-handed?” in Google will produce an A.I.-generated summary of the answer on top of the search results. And whenever you use the search tool inside Instagram, you may now be interacting with Meta’s chatbot, Meta AI. In addition, when Apple’s suite of A.I. tools, Apple Intelligence, arrives on iPhones and other Apple products through software updates this month, the tech will appear inside the buttons we use to edit text and photos.

The proliferation of A.I. in consumer technology has significant implications for our data privacy, because companies are interested in stitching together and analyzing our digital activities, including details inside our photos, messages and web searches, to improve A.I. systems. For users, the tools can simply be an annoyance when they don’t work well.

“There’s a genuine distrust in this stuff, but other than that, it’s a design problem,” said Thorin Klosowski, a privacy and security analyst at the Electronic Frontier Foundation, a digital rights nonprofit, and a former editor at Wirecutter, the reviews site owned by The New York Times. “It’s just ugly and in the way.”

It helps to know how to opt out. After I contacted Microsoft, Meta, Apple and Google, they offered steps to turn off their A.I. tools or data collection, where possible. I’ll walk you through the steps.

Google

Google’s highest-profile A.I. product, A.I. Overviews, automatically generates a summary that tries to answer questions you enter into a Google search. The feature had a rocky debut in May — when, among other snafus, Google’s A.I. told users that they could put glue on pizza — but it has since improved.

Still, the A.I. summaries can be distracting, and there’s no way to deactivate them from loading, but you can click a button to filter them out. After typing something like “chocolate chip cookies recipe” into a search bar, click the “Web” tab to see a list of plain search results, just as Google search used to be.

As for search data, users can prevent Google from keeping a record of their web searches by visiting myactivity.google.com and switching off “web and app activity.”

Google also has an A.I. chatbot, Gemini, and the setting to prevent it from storing data can be found at myactivity.google.com/product/gemini.

Meta

In April, Meta AI, a chatbot that can look up flights, generate images and whip up recipes, began appearing in the search bar of Meta’s apps, including Instagram, WhatsApp and Messenger. There is currently no way for users to turn off Meta AI, Meta said.

Only in regions with stronger data protection laws, including the European Union and Britain, can people deny Meta access to their personal information to build and train Meta’s A.I.

On Instagram, for instance, people living in those places can click on “settings,” then “about” and “privacy policy,” which will lead to opt-out instructions. 

Everyone else, including users in the United States, can visit this support page to ask Meta only to delete data used by third parties to develop its A.I.

Microsoft

Microsoft’s A.I. chatbot, Copilot, can be activated by clicking a rainbow button built into some products like the Edge browser and Bing search.

The simplest way to avoid the chatbot is not to click on that button. But if you want to remove it from the Edge browser, you can enter edge://settings into the address bar and click “Sidebar,” then “App and notification settings” and, finally, “Copilot,” where you should toggle off the Copilot setting.

If you want to prevent Copilot from using your data to train the A.I., you have to visit copilot.microsoft.com and go into the privacy menu in the account settings, where you can toggle off an option labeled “Model training.”

A bonus tip for users of LinkedIn, Microsoft’s social network for professionals: The site recently began using anything posted on its site to train its A.I. system, which could eventually be used to help people find new jobs.

To prevent LinkedIn from using your content, go into the Settings and Privacy tab under your profile, click the “Data privacy” tab and click on “Data for GenAI Improvement.” Then toggle the switch off.

(The Times sued Microsoft and its partner OpenAI last year for using copyrighted news articles without permission to train chatbots.)

Apple

Apple’s suite of A.I. services, Apple Intelligence, will be released this month in an unfinished state through software updates on some iPhones, iPads and Macs. To use Apple Intelligence, users will have to opt in through a menu labeled “Apple Intelligence & Siri.”

Once activated, some of the features will appear inside tools for editing text and photos — when you edit a photo, for instance, there’s a “Clean Up” button to automatically remove photo bombers.

If you change your mind and no longer want to use Apple Intelligence, you can go back into the settings and toggle the Apple Intelligence switch off, which makes the tools go away.

Apple says it has devised a system that protects users privacy, in which data pushed to its servers is inaccessible to Apple. Rather, the company says, it is used exclusively to process a user’s request, such as a complex question posed to Siri, before the information is purged from its servers." [1]

1. How to Say No to Our A.I. Overlords. Chen, Brian X.  New York Times (Online) New York Times Company. Oct 9, 2024.

Komentarų nėra: