The Age of All-Access AI Agent is Here

Years, i the cost of using the “free” services from Google, Facebook, Microsoft, and other Big Tech companies to share your data. Uploading your life to the cloud and using free technology brings convenience, but it puts personal information in the hands of big companies who are often looking to make money from it. Now, the next wave of productive AI systems may want more access to your data than ever before.

Over the past two years, productive AI tools—such as OpenAI’s ChatGPT and Google Gemini—have moved beyond the relatively simple, text-only chatbots that companies initially released. Instead, Big AI is increasingly building and moving toward the adoption of agents and “assistants” that promise to take steps and complete tasks on your behalf. The problem? To get the most out of them, you’ll need to give them access to your systems and data. Although the first big argument about large-scale linguistic models (LLMs) was the obvious copying of copyrighted data on the Internet, the access of AI agents to your personal data can cause many problems.

“AI agents, in order to be able to work fully, to be able to access applications, usually need access to the operating system or the OS level of the device you are using,” said Harry Farmer, a senior researcher at the Ada Lovelace Institute, whose work has included studying the impact of AI assistants and found that they can cause a “significant threat” to Internet security and privacy. Personalizing chatbots or assistants, Farmer says, may involve data sharing. “All those things, in order to work, need a lot of information about you,” he said.

While there is no strict definition of what an AI agent actually is, it is often best thought of as a generative AI system or LLM given a certain degree of autonomy. Currently, agents or assistants, including AI web browsers, can take control of your device and browse the web for you, book flights, conduct research, or add items to shopping carts. Others can complete tasks that involve dozens of individual steps.

Although current AI agents are fragile and often unable to complete tasks assigned to them, technology companies are betting that the systems will dramatically replace millions of human jobs as they become more powerful. The main part of their use may come from accessing data. So, if you want a system that can provide you with your schedule and tasks, it will need access to your calendar, messages, emails, and more.

Some of the more advanced AI products and features offer a glimpse of how much access agents and systems can offer. Certain agents developed for businesses can read code, emails, databases, Slack messages, files stored in Google Drive, and more. Microsoft’s controversial Recall product takes screenshots of your desktop every few seconds, so you can search everything you’ve done on your device. Tinder has created an AI feature that can search through photos on your phone to “better understand” the “interests and personalities” of users.

Carissa Véliz, an author and associate professor at the University of Oxford, says that consumers often have no real way to test how AI or technology companies treat their data in the ways they claim. “These companies behave very badly with data,” said Véliz. “They showed a great lack of respect for privacy.”