Notes I took from the artificial intelligence news that caught my eye between March 16 and March 20.

image credit: Google
  • Google introduced a new artificial intelligence (AI) model called VLOGGER . This model can create an animated video with facial expressions and head movements from a single photo and an audio track. ‘Entertainment’, ‘Education’, ‘Virtual Assistants’, ‘Customer Service’, and ‘Distance Education’ are foreseen as potential use cases for the artificial intelligence application, which is still undergoing trials. Using a large multimedia dataset called MENTOR, the model can create surprisingly realistic videos by using neural network technology to predict motion, body, pose, gaze, and expressions.
  • The fact that surprisingly realistic content can now be produced from even a single photo can turn some people’s lives into a nightmare. Deepfake or deepnude are some of the words we will hear quite often in the future. A news story I saw on Euronews is an example of this. Nude images of a 21-year-old semi-professional Belgian model named Julia were created using deepfake. Julia, who later applied to the prosecutor’s office, was deterred due to the prosecutor’s workload and the low success rate in such crimes. So Julia decided to draw attention to this incident. I have included similar news before. In these days when celebrities’ ‘deepfake porn’ has become commonplace, I think the fact that it has reached a level that can even affect the lives of ordinary people should make us think about taking precautions. I recommend watching the video below that I watched on YouTube yesterday related to this news.
  • Hugging Face is entering the open-source robotics field. It published a job posting for an “Embodied Robotics Engineer” position focused on designing, building, and maintaining open-source and low-cost robotic systems integrating artificial intelligence (especially deep learning and embodied AI). The project aims to develop low-cost and accessible AI robots with off-the-shelf components and 3D printed parts.
  • Is it time to entrust our security to artificial intelligence now? Some are about to start trying. The Las Vegas Metropolitan Police Department (LVMPD) is reportedly planning to use artificial intelligence (AI) to make the Fremont Street Experience safer. The police department, which promises responsible and transparent use despite concerns over freedom and privacy rights, has made plans to monitor the area with AI-supported cameras, intelligent scene analysis tools, and smart streetlights. I genuinely wonder what kind of impact it will have when the system becomes operational.
  • While all major technology companies were offering their artificial intelligence products to consumers, the lack of a move from Apple made me suspicious. The news I saw today has already started writing rumors that Apple is in talks to bring Google’s Gemini AI features to iPhones. I think it is a bit humiliating for the company that boasts of making the most valuable and best phones in the world. ( News 1 , News 2 , News 3 )
  • As Elon Musk promised earlier, X’s artificial intelligence Grok has become open-source. The artificial intelligence Grok, whose Github repo has been published, is a large language model (LLM) trained on a massive text dataset. It is not tailored to a specific task and can be used for a variety of purposes. I think a free and open source is always welcome for developers. On the other hand, whether it is at a level to compete with Gemini and ChatGPT is a question that those who understand this business can answer.
  • YouTube announced its new practice for artificial intelligence (AI) generated content on its platform on its official blog. A new label application will be started for videos created or significantly altered using AI. In the descriptions of such videos, it will be mandatory to include a clear disclosure that the content was created with AI. AI content used for specific purposes such as documentaries, education, or scientific research may be exempt from the label.
  • The most important event of this week, or even this month, was of course the introduction of Nvidia’s Blackwell GPU architecture. The B200 GPUs, which perform up to 5 times better and consume up to 4 times less power than the H100 chips that give life to the artificial intelligence technology we currently use, were introduced. The abbreviation GPU (graphics processing unit) is no longer enough to describe these chips. Because now everyone’s eyes are on artificial intelligence, and these new chips and all other related introduced products will form the backbone of next-generation artificial intelligence. Perhaps with this Blackwell architecture, we will have the opportunity to see first-of-their-kind AGIs (Artificial General Intelligence). All major companies are already lining up for the cards, which will be released later this year. I watched the first hour of the event with my mouth open. It was an event that artificial intelligence enthusiasts must definitely watch. Nvidia continues to put on a show… (Update: Nvidia CEO Jensen Huang said in an interview with CNBC that the prices of Blackwell GPUs are between 30,000 and 40,000 dollars.)
  • Saudi Arabia plans to create a $40 billion fund to encourage innovation in the field of artificial intelligence (AI). This fund will reportedly finance research and development in areas such as data, artificial intelligence, and sustainability programs. They reportedly aim to accelerate skill development and create high-quality job opportunities for young information technology professionals in Saudi Arabia. As followers of this site know, all major companies that want to develop next-generation artificial intelligence chips and invest in technologies in the field of artificial intelligence are knocking on the door of Arab capital. Investment news from other Arab countries in this field may also arrive in the coming years.