automation Archives - The Media Copilot https://mediacopilot.ai/tag/automation/ How AI is changing Media, journalism and content creation Thu, 21 May 2026 23:23:35 +0000 en-US hourly 1 https://wordpress.org/?v=7.0 https://mediacopilot.ai/wp-content/uploads/2024/08/cropped-cropped-Media-Copilot-favicon-60x60.jpeg automation Archives - The Media Copilot https://mediacopilot.ai/tag/automation/ 32 32 Chaos erupts after Anthropic forces Clawdbot rebrand https://mediacopilot.ai/clawdbot-moltbot-rebrand-anthropic-trademark/ Wed, 28 Jan 2026 13:00:00 +0000 https://mediacopilot.ai/?p=3602 Illustration of a giant orange lobster confronting hooded crypto scammers with Bitcoin coins flyingThe viral AI assistant is now Moltbot — but not before crypto scammers hijacked its old accounts.

The post Chaos erupts after Anthropic forces Clawdbot rebrand appeared first on The Media Copilot.

]]>

Peter Steinberger’s viral AI assistant Clawdbot is now called Moltbot after Anthropic issued a trademark request Monday morning. The name “Clawd” was too similar to “Claude,” the AI model that powers many Clawdbot installations.

Key Takeaways

  • Anthropic forced viral Clawdbot to rebrand to Moltbot over the Claude trademark.
  • Crypto scammers grabbed @clawdbot in 10 seconds and pumped $CLAWD to $16M.
  • Shows how brittle creator-account identity is during forced AI rebrands.

“Anthropic asked us to change our name (trademark stuff), and honestly? ‘Molt’ fits perfectly — it’s what lobsters do to grow,” Steinberger wrote on X.

The rebrand execution went sideways fast. During a 10-second window while renaming the project’s X and GitHub accounts, crypto scammers grabbed the abandoned handles. The old @clawdbot accounts now pump token scams to followers who don’t know about the switch.

A fake $CLAWD token hit $16 million market cap before Steinberger publicly denounced it: “I will never do a coin. Any project that lists me as coin owner is a SCAM.”

The chaos comes as security researchers flag concerns about exposed Moltbot instances. Slowmist reported multiple unauthenticated servers publicly accessible, with flaws that “may lead to credential theft and even remote code execution.” Researcher Jamieson O’Reilly found hundreds of misconfigured installations via Shodan.

For newsrooms experimenting with AI agents, the episode highlights two things: how fast trademark conflicts emerge in this space, and why self-hosted tools require careful security hygiene. Running an AI assistant with shell access is powerful — but exposing it to the internet without authentication is asking for trouble.

The official project now lives at molt.bot and github.com/moltbot/moltbot. Steinberger is working with GitHub to recover the hijacked accounts.

The post Chaos erupts after Anthropic forces Clawdbot rebrand appeared first on The Media Copilot.

]]>
Anthropic’s new Cowork tool brings agentic AI to everyday newsroom tasks https://mediacopilot.ai/anthropic-cowork-agentic-ai-knowledge-work/ Tue, 13 Jan 2026 13:00:00 +0000 https://mediacopilot.ai/?p=3314 The feature lets Claude access folders on your Mac to automate reports, expenses and file management. But there are risks.

The post Anthropic’s new Cowork tool brings agentic AI to everyday newsroom tasks appeared first on The Media Copilot.

]]>

Anthropic wants its Claude AI to handle your busywork.

Key Takeaways

  • Cowork extends Claude Code’s agentic capability to general macOS work.
  • Users grant folder access; Claude runs plain-language tasks locally.
  • Folder-level access raises source-confidentiality and security risks.

The company launched Cowork, a new feature that brings the agentic capabilities of its popular Claude Code tool to general knowledge work. Built into the macOS desktop app, Cowork lets users give Claude access to a specific folder and then issue plain language instructions for tasks.

Think filling out expense reports from a folder of receipt photos. Or writing summaries from a stack of interview notes. Or finally cleaning up that chaotic desktop.

For newsrooms drowning in administrative tasks while chasing deadlines, the pitch is compelling. Anthropic says it developed Cowork partly because people were already using Claude Code for general knowledge work tasks anyway.

The key difference from Claude Code: accessibility. While Claude Code required technical know-how to configure, Cowork is designed so any knowledge worker can start immediately. Samuel Axon wrote in Ars Technica that “Anthropic’s goal with Cowork is to make it something any knowledge worker—from developers to marketers—could get rolling with right away.”

But there are real concerns journalists should understand before handing over folder access.

Vague prompts can cause problems. Poorly worded instructions or bad luck can lead the agent to do destructive things like deleting files unexpectedly. And prompt injection attacks remain an unsolved risk, meaning malicious content in documents could potentially manipulate the AI’s behavior.

These aren’t hypothetical concerns. They’re the kind of risks that technical users of Claude Code understood going in. Less technical users might not have that foresight.

For now, Anthropic is moving cautiously. Cowork is available only as a research preview to Max subscribers, with no timeline for wider release.

Why it matters for newsrooms: Agentic AI tools that can work autonomously on file management and document creation could significantly reduce time spent on administrative tasks. But the current limitations and risks mean most newsrooms should wait before deploying this in production workflows. Early adopters should treat it as an experiment, not a solution.

The launch comes alongside Anthropic’s announcement of Claude for Healthcare, as the company expands beyond its developer-focused roots into broader professional markets.

The post Anthropic’s new Cowork tool brings agentic AI to everyday newsroom tasks appeared first on The Media Copilot.

]]>