The Age of the All-Access AI Agent Is Here

For years, the The cost of using “free” services from Google, Facebook, Microsoft and other big tech companies has resulted in your data being handed over. Uploading your life to the cloud and using free technology is convenient, but it puts personal information in the hands of giant corporations who will often seek to monetize it. Now, the next wave of generative AI systems will likely want access to your data more than ever.

Over the past two years, generative AI tools, such as OpenAI’s ChatGPT and Google’s Gemini, have overtaken the relatively simple, text-only chatbots that companies initially launched. Instead, Big AI is increasingly building and pushing toward the adoption of agents and “assistants” that promise they can take actions and complete tasks on your behalf. The problem ? To get the most out of it, you will need to grant them access to your systems and data. While much of the initial controversy over large language models (LLMs) concerned the blatant copying of copyrighted data online, AI agents’ access to your personal data is likely to cause a new set of problems.

“AI agents, to benefit from their full functionality and to be able to access applications, often need access to the operating system or operating system level of the device you’re running them on,” says Harry Farmer, a senior fellow at the Ada Lovelace Institute, whose work has included studying the impact of AI assistants and finding that they can pose a “profound threat” to cybersecurity and privacy. For personalization of chatbots or assistants, Farmer says, there can be data trade-offs. “All of these things, to work, require a lot of information about you,” he says.

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 with some level of autonomy. Currently, agents or assistants, including AI web browsers, can take control of your device and browse the web for you, book flights, perform searches or add items to shopping carts. Some can perform tasks with dozens of individual steps.

While current AI agents are flawed and often can’t accomplish the tasks they’ve been tasked to do, tech companies are betting that the systems will fundamentally change the work of millions of people as they become more capable. A key part of their usefulness likely comes from access to 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 more advanced AI products and capabilities provide insight into the degree of access that could be granted to agents and systems. Some agents developed for businesses can read code, emails, databases, Slack messages, files stored in Google Drive, etc. Microsoft’s controversial Recall product takes screenshots of your desktop every few seconds, so you can look up everything you’ve done on your device. Tinder has created an AI feature that can search photos on your phone “to better understand” users’ “interests and personalities.”

Carissa Véliz, author and associate professor at the University of Oxford, says that most of the time, consumers have no real way to verify whether AI or tech companies are treating their data as they claim. “These companies are very dispersed when it comes to data,” says Véliz. “They showed little respect for privacy.”