AI Daily: DeepSeek banned on U.S. government devices

Catch up on the top artificial intelligence news and commentary by Wall Street analysts on publicly traded companies in the space with this daily recap compiled by The Fly:

DEEPSEEK BANNED: The U.S. Commerce Department is banning Chinese AI model DeepSeek on government devices, according to a message seen by Reuters and two people familiar with the matter, Karen Freifeld reports. “To help keep Department of Commerce information systems safe, access to the new Chinese based AI DeepSeek is broadly prohibited on all GFE,” said one mass email to staffers about their government-furnished devices. DeepSeek is a Chinese-built large-language open-source model that claims to rival offerings from Microsoft-backed (MSFT) OpenAI’s ChatGPT and Meta Platforms (META) but using a much smaller budget.

LLAMA AI MODEL: Meta said in a statement, “Our open source AI model, Llama, has been downloaded more than one billion times. We believe open sourcing AI models is essential to ensuring people everywhere have access to the benefits of AI, and every download of Llama moves us closer to that goal. Among those one billion downloads are industry leader Spotify (SPOT), startup Fynopsis and enterprising developer Srimoyee Mukhopadhyay.”

INVESTING HEAVILY: In an interview on CNBC’s Mad Money, Intuit’s (INTU) CEO Sasan Goodarzi said AI has been a big part of Intuit’s innovation. “With the investments we’ve made, we’re being disruptive.” Accountants are big partners, especially for larger firms, he noted. Most of the companies Intuit works with are service-based not product-based, so tariff talks are not having a big impact on them, he added.

AI RECOMMENDATIONS: Knightscope (KSCP) has formally submitted public comment to the White House’s “Development of an Artificial Intelligence Action Plan” emphasizing the urgent need for AI-driven security solutions to safeguard federal facilities and critical infrastructure, while also fostering safer environments for government employees, visitors, and communities. This policy initiative, led by the Office of Science and Technology Policy under the new administration, seeks to strengthen America’s AI capabilities while removing barriers to private sector innovation. Knightscope’s recommendations urge the federal government to modernize outdated procurement policies and accelerate the deployment of autonomous security robots in military bases, government buildings, and high-risk environments. The company proposes key initiatives, including authorizing autonomous security in federally managed spaces, launching a multi-agency pilot program, and ensuring national security by prohibiting the federal procurement of ASRs from adversary nations.

SMART SHOP: Instacart (CART) introduced Smart Shop, which leverages cutting-edge generative AI and machine learning to create a more personalized grocery shopping experience. With Smart Shop, Instacart is making online grocery shopping more intuitive by analyzing customer habits and dietary preferences to surface the most relevant products faster. In addition to Smart Shop, Instacart unveiled new AI-powered Health Tags, which provide detailed and transparent nutritional information across the catalog, and Inspiration Pages, curated destinations within the Instacart experience featuring expert-backed health recommendations and shoppable recipes. The first Inspiration Page was developed in collaboration with the American Diabetes Association and features evidence-based nutrition guidance and diabetes-friendly grocery and recipe recommendations. Launched to coincide with National Nutrition Month, these offerings make it easier for consumers to discover relevant products based on their unique health and lifestyle preferences on Instacart.

GENAI ASSISTANT: Verizon (VZ) said, “Small businesses can now harness the power of Generative Artificial Intelligence to supercharge their operations. The newest product built directly for small businesses, Verizon Business Assistant is a GenAI-powered text messaging solution designed to help small businesses automate customer interactions and enhance engagement. Business Assistant provides instant text responses to commonly asked questions, learns and improves over time, and enables businesses to offer faster and more efficient customer service. How it works: If a customer texts a bakery to ask about gluten-free options, Business Assistant can instantly respond using GenAI. If it doesn’t yet have the answer, it will then connect to a live employee for assistance. The best part is that it learns from interactions, building a knowledge base that improves overtime. Ultimately, Business Assistant reduces the need for human intervention, allowing employees to focus on higher-value tasks while ensuring customers get fast, accurate responses.”

PARTNERSHIPS: Booz Allen Hamilton (BAH) and Shield AI announced that the companies are partnering to deliver a new era of AI-enabled, software-defined autonomous solutions for the U.S. government. The partnership will combine Booz Allen’s deep mission engineering expertise with Shield AI’s Hivemind Enterprise platform, initially focused on building solutions for the Department of Defense. The resulting collaboration will accelerate the development and integration of advanced AI control systems and autonomy into a wide variety of uncrewed systems, enabling their rapid deployment into threat scenarios to give the U.S. military essential decision advantage. The capabilities will give armed forces an edge over adversaries and help protect troops in complex, dynamic operational environments. Shield AI’s Hivemind software can pilot autonomous vehicles ranging from one-way attack drones to F-16s, and support swarming operations – even when GPS and communications are jammed. The technological collaboration also has potential applications across other high-priority national security missions.

Akamai (AKAM) announced it has entered into an agreement with Vast Data to make data-intensive AI inferencing faster and more efficient. The collaboration will marry Akamai’s distributed platform with the Vast Data Platform’s approach to data-intensive computing to integrate data for large, distributed customers that need better local response time and localization for distributed inference.