AI News from April 28, 2026

Chinese companies are embedding AI into cars, robots, and industrial products as the technology shifts toward edge devices.

DeepSeek’s latest model improved performance but did not deliver a major market jolt in a rapidly shifting AI race.

Activist investor Starboard Value reportedly built a position in Dynatrace, a software company positioned around AI operations.

EU regulators issued guidance aimed at ensuring rival AI services can access Google’s Android ecosystem.

Taylor Swift’s team filed trademark applications aimed at protecting her voice and image amid AI deepfake concerns.

A jury was seated as Elon Musk’s lawsuit against Sam Altman and OpenAI moved into trial proceedings.

Google reportedly agreed to let the Pentagon use its AI models for classified work despite internal objections.

Reports say OpenAI missed internal growth and revenue goals, raising questions about AI infrastructure spending.

Chinese AI models from DeepSeek, Qwen, and Moonshot are cheaper, more adaptable, and nearly as capable as leading US platforms, raising competitive concerns.

Bridgewater CIOs warn that AI is threatening legacy software firms, with markets already pricing in disruption risk for application software companies.

South Africa pulled its draft national AI policy after an investigation revealed the document cited sources fabricated by an AI chatbot.

Analyst Ming-Chi Kuo reports OpenAI is developing a smartphone with AI agents replacing apps, in collaboration with Qualcomm, MediaTek, and Luxshare.

More than 600 Google staff signed a letter asking CEO Sundar Pichai to block the Pentagon from using Google AI models for classified military purposes.

OpenAI and Microsoft announced a revised deal dropping Microsoft's exclusive right to sell OpenAI models and capping revenue share payments to the tech giant.