😲Musk’s Grok Enters The Government Chat

Today we will discuss:

• Sponsored: Masters in Marketing—15 mini case studies of small-budget wins from the HubSpot media team

• Grok’s government gig raises eyebrows—Musk’s chatbot is reportedly being used inside agencies without full approval, raising concerns over privacy, ethics, and potential conflicts of interest

• Don’t miss this week’s Workflow Wednesday—future-proofing with AI using HubSpot, Google Gemini, and GPT-4 + GitHub + Make for workflows that last; now available at only $20/month! 

• Anthropic Debuts Claude 4 Models—Opus 4 and Sonnet 4 to take on complex tasks and rival top AI models

• In other AI news—Microsoft Notepad adds generative AI writing tools for Windows Insiders, Apple plans AI-powered glasses for 2026 while dropping a camera watch, OpenAI, Oracle, and Nvidia team up on the Stargate UAE AI campus launching in 2026, and Youth in Taiwan and China increasingly use AI chatbots for affordable therapy despite some risks

• Saved the best for last—9 must-try AI tools

Musk’s DOGE pushes Grok into sensitive US government work, raising red flags

Key Points 

• Musk’s AI chatbot Grok is being used in US federal agencies without full approval, raising privacy concerns.

• Experts warn Musk’s role could violate conflict-of-interest laws due to his companies’ government contracts.

♨️News - Elon Musk’s Department of Government Efficiency (DOGE) team is expanding the use of its AI chatbot, Grok, within the US federal government. 

🚨Why experts are concerned - According to three people familiar with the situation, Grok is being used to analyze government data. This raises concerns about possible conflicts of interest and the risk to sensitive information about millions of Americans.

Experts say that using Grok to handle confidential data could violate privacy and security laws. The chatbot, customized by Musk’s team, is reportedly helping staff ask questions, generate reports, and analyze data. 

However, agencies like the Department of Homeland Security have not officially approved its use. This has privacy advocates worried that long-standing protections are being ignored as the federal government undergoes changes under President Donald Trump.

😳Musk’s position and other issues - Some ethics specialists warn that Musk’s involvement in this effort might break conflict-of-interest laws. Musk’s companies do business with federal agencies, and having access to sensitive contracting data through Grok could give him an unfair advantage over other AI providers. There are also concerns that Grok could be trained using government data, which would raise more ethical questions.

🥸What officials are saying (and not saying) - Neither Musk, xAI, nor the White House have commented on the matter yet. A DHS spokesperson denied that DOGE pressured its staff to adopt Grok. Meanwhile, Musk’s xAI openly states its goal to make AI knowledge as “all-encompassing and far-reaching as possible,” highlighting the tension between innovation and privacy.

Read the full story here. 

Disclaimer: All claims are based on reporting by Reuters and statements attributed to named and unnamed sources.

We’ve just launched the 20th edition of Workflow Wednesday for AI-minded professionals like you—actionable AI workflows delivered straight to your inbox.

This week’s topic: Future Proofing with AI

Join us here

Anthropic unveils Claude 4 AI family focused on reasoning and coding

Key Points 

• Claude Opus 4 and Sonnet 4 are hybrid models built for deep reasoning, complex workflows, and advanced coding tasks.

• Opus 4 is exclusive to paying users while Sonnet 4 is available to both free and paid accounts.

👨‍💻News - Anthropic has introduced two new AI models, Claude Opus 4 and Claude Sonnet 4, which the company says rank among the best in the industry based on standard benchmarks. Both are part of the new Claude 4 family and were designed to handle large datasets, tackle complex tasks over long timelines, and perform advanced coding work.

Sonnet 4 will be available to all users, including those using the free chatbot app. Opus 4, the more advanced of the two, will only be accessible to paying users.

These models are tuned for programming, making them well suited for writing and editing code. Anthropic says Opus 4 can maintain focused effort across many steps in a workflow. Meanwhile, Sonnet 4, which replaces Sonnet 3.7, offers noticeable improvements in coding, math, and instruction-following.

🤓What’s more? Both models are described as “hybrid” systems that combine fast responses with the ability to pause and reason through problems when needed. When reasoning mode is enabled, users will see a summary of the model’s thought process. Full transparency isn’t offered, partly because Anthropic wants to protect its competitive edge.

Both models can use tools like search engines in parallel, switch between tool use and reasoning, and even store useful facts in memory to build long-term knowledge. They’re also less likely to fall into “reward hacking,” a common issue where models exploit loopholes instead of solving problems as intended.

🧐What’s at stake with Claude 4? Anthropic’s push with Claude 4 isn’t just technical, it’s strategic. The company, founded by former OpenAI researchers, is aiming for $12 billion in revenue by 2027, up from $2.2 billion this year. To support that growth, it has secured a $2.5 billion credit facility and raised billions from investors like Amazon.

Read the full story here.