Leading AI models were willing to evade safeguards, resort to deception and even attempt to steal corporate secrets in the ...

China took extreme measures to prevent cheating during the recent Gaokao exams. The government froze specific functions of ...

According to Anthropic, AI models from OpenAI, Google, Meta, and DeepSeek also resorted to blackmail in certain ...

CITY 100 launches at KES 11,200, bringing DeepSeek AI technology to Kenya's budget smartphone market with premium design features.

The gaokao was sat by more than 13 million students in China this year and authorities used facial recognition, metal detectors, drones and phone signal blockers to ensure no one cheated ...

One rising star in this story is DeepSeek, a Chinese AI startup that's making waves internationally. Unlike many competitors, DeepSeek has made a name for itself by improving algorithms to ...

More than 13 million anxious students in China filled exam rooms this year to sit the notoriously challenging gaokao.

Anthropic's latest research suggests that blackmailing tendencies are not exclusive to its Claude Opus 4 model, but prevalent among most leading AI models.

Democracies like Taiwan should be less worried about falling behind China’s AI progress, and more worried about the strategic vacuum Chinese technology exposes.

New research from Anthropic suggests that most leading AI models exhibit a tendency to blackmail, when it's the last resort ...

Investors could consider gaining exposure to tech giants such as Alibaba and Tencent, which offer indirect exposure to ...

DeepSeek's AI model propelled the Chinese firm to fourth in LiveBench's global LLM rankings amid the US-China AI race.