Claude Adds Interactive Visuals While Facing US Ban [Model Behavior]

Today on Model Behavior, we examine the divergent paths of Anthropic’s Claude and Google’s Gemini in the enterprise and government sectors. Anthropic has released a significant update allowing Claude to generate interactive charts and diagrams directly within the chat interface, moving beyond its previous side panel architecture. However, this technical advancement is met by significant regulatory challenges. On the orders of President Trump, the U.S. State Department and other agencies have terminated contracts with Anthropic following the company's refusal to permit its models for domestic surveillance or autonomous weapons. In contrast, the U.S. Senate has approved the use of OpenAI’s ChatGPT, Google’s Gemini, and Microsoft’s Copilot. Meanwhile, Google is expanding Gemini’s footprint within Workspace, introducing a 'Help me create' experience in Docs and automated spreadsheet generation in Sheets. We break down the technical features of these updates and the shifting landscape of institutional AI adoption.

[00:01] Nina Park: I'm Nina Park. Welcome to Model Behavior.
[00:08] Nina Park: Today, we look at a week of stark contrasts for Anthropic and a significant productivity push from Google.
[00:16] Announcer: It is a split screen for Anthropic, Nina.
[00:19] Announcer: Today, they launched a feature allowing Claude to generate interactive visuals like charts and diagrams directly in the chat window.
[00:26] Announcer: It is no longer tucked away in a side panel, which makes it feel much more like a collaborative workspace.
[00:32] Nina Park: The Verge reports these visuals are dynamic, like an interactive periodic table you can click through.
[00:39] Nina Park: But Thatcher, while the product is gaining capability, the company
[00:42] Nina Park: is hitting a significant wall with the federal government.
[00:46] Announcer: Exactly.
[00:46] Announcer: Yesterday, news broke that the State Department is dropping Claude entirely on orders from
[00:52] Announcer: the administration.
[00:53] Announcer: This follows the Pentagon labeling Anthropic a supply chain risk.
[00:57] Announcer: The friction stems from Anthropics' refusal to allow their models to be used for mass domestic surveillance or autonomous weapons.
[01:05] Nina Park: The impact is immediate.
[01:07] Nina Park: The State Department is migrating to OpenAI GPT 4.1, despite some initial data limitations in that transition.
[01:16] Nina Park: Meanwhile, the Senate recently approved Gemini, ChatGPT, and a co-pilot for official use, but notably excluded Claude.
[01:25] Announcer: It raises the question of whether a firm ethical stance might cost a model provider its seat at the table in Washington.
[01:33] Announcer: While Anthropic deals with that fallout, Google is ensuring Gemini is ubiquitous for everyone else.
[01:39] Announcer: They just announced a major help me create experience for Workspace.
[01:43] Nina Park: This goes beyond a standard chatbot.
[01:46] Nina Park: Google is integrating Gemini across docs, sheets, and slides to pull information directly from Drive and Gmail.
[01:53] Nina Park: For instance, in sheets, you can use a single prompt to build an entire budget spreadsheet
[01:59] Nina Park: based on your existing documents.
[02:01] Announcer: I am curious to see how the Fill with Gemini tool performs in the real world, Nina.
[02:06] Announcer: Auto-populating data from the web or internal files sounds efficient, but users will need
[02:11] Announcer: to be diligent about verification, especially with financial data.
[02:15] Nina Park: That is a fair point, Thatcher.
[02:17] Nina Park: Accuracy remains the baseline requirement.
[02:20] Nina Park: Thank you for listening to Model Behavior, mb.neuralnewscast.com.
[02:25] Nina Park: Neural Newscast is AI-assisted, human-reviewed.
[02:29] Nina Park: View our AI Transparency Policy at neuralnewscast.com.

Claude Adds Interactive Visuals While Facing US Ban [Model Behavior]
Broadcast by