AI agents could make the internet go dark, top analysts warn

Will agentic middlemen lurk in the shadows of what was once called cyberspace?

Financial analysts warn that widespread adoption of AI agents could fundamentally transform how humans interact with digital services and information online. These AI-powered personal assistants, which can perform tasks across various applications on behalf of users, threaten to disrupt existing business models and potentially reshape the economics of the internet. The implications extend beyond simple convenience features to potentially affecting how news is consumed, services are marketed, and digital advertising functions.

The big picture: Analysts at Bernstein Research predict that if AI agents become truly useful, they could dramatically alter how users interact with online services and information, effectively making the “internet go dark.”

• The shift would move direct human interaction away from apps and websites as AI agents become intermediaries for tasks and information gathering.
• This transformation could fundamentally change how digital services are designed, moving from human-centric interfaces to AI-optimized systems.

Key implications: The rise of AI agents could disrupt existing business models and revenue streams across the digital economy.

• News organizations could lose advertising revenue if AI agents simply summarize their content without users visiting their websites.
• Ride-sharing platforms might lose direct customer relationships as AI agents automatically select the cheapest available option.

Behind the numbers: The consolidation of user demand through AI agents could create new power dynamics in the digital marketplace.

• Current aggregators who control supply could lose influence as AI agents become the primary interface between consumers and brand services, acting as the ultimate middleman
• The shift could give unprecedented economic power to AI agent creators who would become the new gatekeepers of consumer choice.

Counterpoints: Several significant challenges and uncertainties remain regarding the widespread adoption of AI agents.

• Current AI technology’s accuracy and reasoning abilities often fall short of marketing hype.
• Questions persist about AI agents’ reliability in high-stakes industries like healthcare and insurance.

Where we go from here: According to Bernstein analysts Devnani and Shmulik, the emergence of “Agentic AI” appears inevitable, though its exact impact remains uncertain.

• Companies will need to adapt their business models to account for AI-mediated customer interactions.
• The transformation could accelerate existing concerns about AI’s impact on traditional human employment and economic structures.