Microsoft identifies four cybercriminals behind Storm-2139, an AI abuse scheme exploiting Azure OpenAI services for illicit content.
Category: robotics/AI – Page 76
A dataset used to train large language models (LLMs) has been found to contain nearly 12,000 live secrets, which allow for successful authentication.
The findings once again highlight how hard-coded credentials pose a severe security risk to users and organizations alike, not to mention compounding the problem when LLMs end up suggesting insecure coding practices to their users.
Truffle Security said it downloaded a December 2024 archive from Common Crawl, which maintains a free, open repository of web crawl data. The massive dataset contains over 250 billion pages spanning 18 years.
A new variant of the Vo1d malware botnet has grown to 1,590,299 infected Android TV devices across 226 countries, recruiting devices as part of anonymous proxy server networks.
This is according to an investigation by Xlab, which has been tracking the new campaign since last November, reporting that the botnet peaked on January 14, 2025, and currently has 800,000 active bots.
In September 2024, Dr. Web antivirus researchers found 1.3 million devices across 200 countries compromised by Vo1d malware via an unknown infection vector.
Microsoft has named multiple threat actors part of a cybercrime gang accused of developing malicious tools capable of bypassing generative AI guardrails to generate celebrity deepfakes and other illicit content.
An updated complaint identifies the individuals as Arian Yadegarnia from Iran (aka ‘Fiz’), Alan Krysiak of the United Kingdom (aka ‘Drago’), Ricky Yuen from Hong Kong, China (aka ‘cg-dot’), and Phát Phùng Tấn of Vietnam (aka ‘Asakuri’).
As the company explained today, these threat actors are key members of a global cybercrime gang that it tracks as Storm-2139.
As we listen to speech, our brains actively compute the meaning of individual words. Inspired by the success of large language models (LLMs), we hypothesized that the brain employs vectorial coding principles, such that meaning is reflected in distributed activity of single neurons. We recorded responses of hundreds of neurons in the human hippocampus, which has a well-established role in semantic coding, while participants listened to narrative speech. We find encoding of contextual word meaning in the simultaneous activity of neurons whose individual selectivities span multiple unrelated semantic categories. Like embedding vectors in semantic models, distance between neural population responses correlates with semantic distance; however, this effect was only observed in contextual embedding models (like BERT) and was reversed in non-contextual embedding models (like Word2Vec), suggesting that the semantic distance effect depends critically on contextualization. Moreover, for the subset of highly semantically similar words, even contextual embedders showed an inverse correlation between semantic and neural distances; we attribute this pattern to the noise-mitigating benefits of contrastive coding. Finally, in further support for the critical role of context, we find that range of neural responses covaries with lexical polysemy. Ultimately, these results support the hypothesis that semantic coding in the hippocampus follows vectorial principles.
The authors have declared no competing interest.
Understanding AI agents with Anthropic’s playbook: what they are, what they can do for your business, and how to set one up. Automate your way to success and freedom.
Kaiming He, a professor in the Department of Electrical Engineering and Computer Science, believes AI can create a common language that lowers barriers between scientific fields and fosters collaboration across scientific disciplines.
“There is no way I could ever understand high-energy physics, chemistry, or the frontier of biology research, but now we are seeing something that can help us to break these walls,” said He.
MIT Associate Professor Kaiming He discusses the role of AI in interdisciplinary collaborations, connecting basic science to artificial intelligence, machine learning, and neural networks.
The question is, can DEI proponents, who are already being marginalized, retool? Can they see themselves as champions who will guide humanity — regardless of peoples’ race, class, sexual orientation, gender, etc. — in this Fourth Industrial Revolution?
For, if political leaders are as unable as they seem to establish meaningful guardrails, AI will push those struggling to live their best lives (a right that should belong to all) to be thrown so far under the bus that roadkill will be more recognizable.
A robotic labor force.
Engineering companies in China have begun the process of starting full-scale mass production of humanoid robots in anticipation of a commercial boom in 2025. Companies such as Shanghai Zhiyuan Innovation Technology (AgiBot) and Shanghai Kepler Robot Company have already been testing the robots to optimize performance while reducing production costs.
With Google’s latest AI model, Gemini 2.0, launched in December, the company is making big moves in AI. Brin’s message is clear, the race to AGI is on, and Google plans to win it.