AI is no longer confined to chatbots and labs. It is seeping into universities, cybersecurity and the very fabric of the internet, and two developments this week show both the scale of that shift and the risks that come with it.
The first concerns what we see online. A new report on Meta's latest AI push suggests the company could accelerate machine-generated content at a scale that reshapes the web itself. The trend is already measurable: researchers estimate that 31.1% of filtered web tokens in August 2026 were AI-generated, up from 10% in June 2024.
The implications go beyond volume. When synthetic text, images and video become hard to tell apart from human work, trust erodes. For a platform built on engagement, cheap content at infinite scale is commercially attractive, but it risks crowding out authentic voices and making credibility the web's scarcest resource.
The second development shows AI expertise becoming a prize in itself. Chinese hackers impersonated a former senior White House technology official, a former State Department economist and a senior Anthropic employee in attempts to break into the cloud accounts of AI policy experts, according to Proofpoint.
The tactic was deceptively simple. Harmless-looking emails, such as invitations to join an "AI Policy Advisory Committee", lured targets, who were then directed to a fake login page built to steal credentials. Proofpoint said targets included experts working on AI regulation, export controls and national AI strategy. Beijing regularly denies US hacking allegations.
The choice of targets is telling. Spies are no longer just chasing code or chips; they want insight into how rival nations plan to govern, restrict and deploy AI.
Together, these stories reveal a single truth: as AI grows more powerful, both information and the people who shape it become more vulnerable.
