After a viral trend showed people using AI coding agents to remove ads and bloatware from smart TVs, some users are reporting that the tools are now refusing to help.
The complaints have focused largely on Anthropic’s Claude, though OpenAI’s Codex has also reportedly blocked some requests. But there’s little evidence that either company deliberately restricted the trend/use case.
Why This Matters
Last week, we covered how people were using Claude Code to strip ads and preinstalled apps from smart TVs raising existential questions about the future of connected TV advertising. (Though, admittedly, the number of people who will actually “reboot” their smart TVs is low.)
Now, some users say Claude is refusing requests to root or reverse-engineer their TVs, sometimes stopping midway through the process.
Importantly, disabling apps through Android’s built-in developer tools is different from rooting a TV or bypassing security protections. The latter appears more likely to trigger safety restrictions or violate the consumer TOS.
Experts React
The reports caught the attention of software developer Eric S. Raymond, who publicly called out OpenAI and Anthropic over rumors that their models were being restricted from helping people remove unwanted “spamware and spyware” from their TVs.
Raymond warned that blocking legitimate uses could drive paying customers toward open-source AI alternatives. He also criticized restrictions that appear to protect corporate interests.
OpenAI’s roon (this is a real person who works at OpenAI for the uninitiated) responded that the blocks are “mostly accidental and a result of difficult decision boundaries.” He added that “there is no prohibition against reverse engineering with codex. it’s getting fixed.”
Meanwhile, other users have shared examples of Claude refusing TV mod requests.
Anthropic has not publicly addressed the reports.
Our Take
The original trend highlights how AI could make ad blocking on CTV more accessible to everyday consumers. It also raises a different question: Who decides what consumers can modify on devices they own? (This is a big debate in other verticals!)
There’s a legitimate distinction between protecting devices from harmful exploits and helping consumers remove unwanted ads or apps. It is a complicated convo, which AI is forcing. This happens!
FWIW, as smart TV makers put more and more (poor quality) ads in front of viewers, it’s not hard to see why more people want the ability to remove them.
