AI News Roundup – Anthropic releases latest Claude model, FAA courts Palantir & others to modernize air traffic control systems, AI companies chart a course for emotionally intelligent chatbots, and more

To help you stay on top of the latest news, our AI practice group has compiled a roundup of the developments we are following. 

• Anthropic has unveiled the latest version of its Claude AI chatbot, as reported by CNBC. Claude Opus 4.7 is, according to Anthropic, “a notable improvement on Opus 4.6 in advanced software engineering.” The company also reported improved performance in vision tasks and working in the real world. However, Opus 4.7’s cybersecurity capabilities are not as advanced as Anthropic’s “Mythos” model, which the company released to a select group last week (as this AI Roundup covered). Instead, Opus 4.7 was released “with safeguards that automatically detect and block requests that indicate prohibited or high-risk cybersecurity uses” that will help future development of Mythos-type models. Anthropic has positioned itself as the AI company most focused on creating safe and responsible AI systems, which has brought it into conflict with the Trump administration as it continues its legal battle with the Department of Defense, but the company appears to be staying the course on opposing “high-risk” AI use. The new model was launched on April 16 and is available via all of Anthropic’s products.

• Bloomberg reports that the U.S. government is soliciting bids for AI tools to aid in air traffic control. The Federal Aviation Administration (FAA) has contacted Peter Thiel’s Palantir Technologies, among other companies, to compete in developing AI-based software to help modernize aging systems. In a statement, the FAA said that it aimed “to revolutionize our air traffic management by modernizing our software, moving to a centralized cloud-based software system that will use artificial intelligence to predict air traffic flows and adjust departure times to resolve conflicts.” The technology may also be used to alert air traffic controllers when aircraft are too close to one another, allowing the controllers to direct the planes to avoid incidents. The FAA has received $12.5 billion from Congress for the modernization initiative but is asking for further funding to complete the project. Further details on the project have not yet been ascertained.

• AI startups are racing to imbue their chatbots with emotional intelligence skills, according to The Atlantic. Ever since the generative AI boom began in autumn 2022 with the release of OpenAI’s ChatGPT, which could converse with users through text, the quality and tone of the model’s output has often affected user perceptions of the technology. While such chatbots existed in the past, the language in ChatGPT was much more natural than past chatbot technologies. Despite this, most chatbots do not focus on emotional preferences, with many AI companies focusing on reasoning abilities or “agentic” AI that can perform tasks for users. A growing proportion of AI chatbot conversations involve emotional or personal issues for users, and companies are beginning to focus on bringing empathy into its models. One startup, Amotions AI based in San Francisco, claims that its model can give tips based on a video input, responding to the person’s tone and facial expressions and adjusting its output accordingly. However, emotional intelligence is difficult to measure even in humans, so evaluating such capabilities in AI is still a difficult challenge for AI labs to face as they seek to implement emotional response into their AI systems.

• The MIT Technology Review covers the 2026 AI Index, the latest edition of the annual report on the state of AI. Stanford University’s Institute for Human-Centered Artificial Intelligence produced the report, now in its sixth year. The Technology Review identified several key takeaways from the report, including the state of the U.S.-China AI rivalry and impacts to jobs. The Index concluded that the U.S. and China are nearly tied, according to the Arena AI benchmark, with models from China’s Alibaba and DeepSeek trailing those from American companies such as Anthropic, OpenAI, and Google DeepMind by very slim margins. However, the report noted that while the U.S. leads in model performance and data center investment, China has the advantage in research output for AI, including publications and patents. The report also covered AI’s impact on jobs, and noted that nearly nine in 10 organizations are using AI in some fashion, while AI may be impacting employment for younger software developers. Many employers also expect to shrink their workforces due to AI in the coming year. For further reading, the full report is available here.

• The UK is launching a $675 million AI investment fund, according to a report from WIRED. The fund, called Sovereign AI, will support the country’s AI startups, who will also gain access to computing resources, visas for hiring international talent, and advice from government AI specialists. The UK government is hoping to reduce dependence on foreign technology and to use AI to stimulate a stagnant economy. Liz Kendall, the UK’s technology secretary, said in a statement that the fund “is unlike anything Government has ever done before. Its unique approach will help break down the barriers that have too often held back British enterprise and innovation.” Some experts doubt that any country could truly become self-sufficient in the field of AI, and that the Sovereign AI fund’s resources are quite small compared to those invested by large American companies, but that smart investment could still help the UK become a leader in certain AI niches.