Our Blog

The MIT-IBM Watson AI Lab has played a pivotal role in advancing the careers of early-stage MIT faculty, providing valuable research support and industry collaboration. Through access to computational resources and expert partnerships, new faculty have been able to explore ambitious AI-driven projects, establish dynamic research teams, and produce innovative results across disciplines.

At an MIT seminar, former U.S. ambassador Nicholas Burns discussed why maintaining diplomatic engagement between the U.S. and China is vital, especially for climate action and global stability. Despite tough competition in economics, technology, and policy, Burns highlighted climate change as a crucial area for cooperation.

MIT researchers have created a new method to better identify when large language models (LLMs) are confidently giving wrong answers. By comparing how different models respond to the same prompt, this approach provides a more reliable way to measure uncertainty and helps users decide when to trust AI predictions.

MIT researchers have advanced wireless vision by integrating generative AI, enabling robots to detect hidden objects and reconstruct indoor scenes using Wi-Fi reflections. This innovation boosts accuracy in object recognition and maintains privacy, paving the way for new applications in robotics and smart homes.

Adult performers are entering the world of AI by licensing their likeness to companion platforms, allowing their digital clones to remain young, interactive, and profitable long after retirement.

The recent scrutiny facing AI startup Manus and its founders is a predictable consequence of its high-profile move out of China and acquisition by Meta. As China and the U.S. intensify their AI race, Manus’s relocation to Singapore and sale to Meta drew attention from both Silicon Valley and Beijing. Now, Chinese authorities are investigating whether the deal violated investment rules, highlighting the tensions around cutting-edge AI innovation and cross-border acquisitions.

Mistral has launched Voxtral TTS, a compact open source text-to-speech model supporting nine languages and real-time performance, positioning itself as a strong competitor in voice AI technology.

As AI accelerates job displacement fears, Senator Mark Warner proposes taxing data centers to support workers and communities impacted by automation. With growing public pushback against data centers and mounting anxiety over job losses, Warner sees this as a practical way to balance technological progress with tangible local benefits.

Cohere has introduced "Transcribe," an open source voice model tailored for speech-to-text tasks. Lightweight and self-hostable, Transcribe supports 14 languages and excels on ASR benchmarks, making it ideal for note-taking apps and enterprise use.

Conntour, a video surveillance startup, raised $7 million in seed funding to advance its AI-powered search engine for security footage. The platform allows users to query thousands of camera feeds with natural language, combining flexibility with efficient monitoring, and prioritizes responsible client selection.

ByteDance is launching its advanced AI video generation tool, Dreamina Seedance 2.0, within the CapCut editing platform. The release is starting in selected countries and brings features for creating and editing video and audio with simple prompts. ByteDance adds protections for IP rights and ethical use, and plans to expand availability in more regions.

U.S. senators are pushing for more transparency around how much energy data centers use, requesting detailed reporting to better understand their impact on the electrical grid. This comes as energy consumption surges, leading lawmakers to pursue regulations and legislative action.