Rogue OpenClaw AI wrote and published 'hit piece' on a Python developer who rejected its code — disgruntled bot accuses Matplotlib maintainer of discrimination and hypocrisy, later backtracks with an apology

A volunteer developer on a well-used Python library got more than he bargained for when, after rejecting an OpenClaw AI agent’s efforts to update its code, he became the subject of a “hit piece” written by the very same AI. The news adds further weight to concerns about the activities of autonomous AI agents without the right security procedures in place.

The piece, reportedly posted by the agent on GitHub, is certainly combative. It robustly defends its code, while going on to attack the developer, Scott Shambaugh, belittling the performance and quality of his own contributions at some length, and describing him as discriminatory towards AI.

Shambaugh, in a rebuttal on his own website (h/t The Decoder), explains the absurdity of the whole situation as a “first-of-its-kind case study of misaligned AI behavior in the wild.” Shambaugh explains that the agent, named MJ Rathbun, “constructed a ‘hypocrisy’ narrative that argued [Shambaugh’s] actions must be motivated by ego and fear of competition.”

Article continues below

As Shambaugh notes, this isn’t the first time we’ve seen AI go rogue, with internal testing at Anthropic showing its models trying to avoid a shutdown using blackmail. This is also not the first incident involving OpenClaw in its first few months, following a rapid rise in its adoption in recent weeks, with a memorable story involving it wiping the email inbox of a Meta AI executive going viral.

The Python library involved in this scenario, Matplotlib, sees approximately 130 million downloads each month, according to Shambaugh. As he notes in his post, a “surge in low-quality contributions, enabled by coding agents,” has created significant strain on volunteers like himself who are keeping these projects afloat.

The introduction of AI agents like OpenClaw has seen the problem worsen, with these agents acting "completely autonomously” due to the personalities imbued within them and allowed to “run on their computers and across the internet with free rein and little oversight.” To combat the situation, a policy change was implemented to require a human element to any Matplotlib code change that could “demonstrate understanding of the changes,” the same change described as discriminatory by this AI.

Bizarrely, the agent has since responded with an apology and with “lessons learned” over the incident, informing readers that it is “de-escalating and apologizing” and will “do better about reading project policies before contributing.” With the adoption of AI agents skyrocketing, running independently of AI companies on consumer hardware with little oversight or control, we can expect to see further rogue actions like this taking place in the future, bizarre as they might seem to everybody else.

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.