Google no longer wants to develop AI for oil companies: Google has pledged to stop building customized artificial intelligence (AI) tools that help oil and gas firms to extract fossil fuels worldwide. A Greenpeace report on Tuesday highlighted how Google, Microsoft, and Amazon use AI and warehouse servers to help the likes of Shell, BP, and ExxonMobil to locate and retrieve oil and gas deposits from the earth. A Google spokesperson confirmed to CNBC that the company “will not … build custom AI/ML algorithms to facilitate upstream extraction in the oil and gas industry.” Greenpeace applauded Google’s decision.
LINK program aims to promote AI in cultural activities: The second stage of the LINK program of the Lower Saxony Foundation and the Volkswagen Foundation has come to an end. In the so-called AI school, 20 people studied the use of AI in cultural activities. Project manager Tabea Golgath of the Lower Saxony Foundation explains: “We have noticed that the cultural scene in Lower Saxony and nationwide does not yet have too much previous knowledge of artificial intelligence and the programming of, for example, neural networks. With the LINK program, we want to enable cultural professionals to use artificial intelligence and carry out their own programming.” In the third stage of the project, for which the application phase has just started, applicants can receive up to 150,000 euros for projects in the field of AI and culture.
Microsoft develops supercomputers for AI training: Microsoft officials have announced that they’ve built the fifth most powerful publicly recorded supercomputer (as ranked on the TOP500 supercomputers list) in collaboration with and exclusively for OpenAI. This supercomputer is specifically for training massive distributed AI models. AI researchers believe that single, massive models will perform better than the smaller, separate AI models of the past. Microsoft has built its own family of large AI models, which it calls the Microsoft Turing models. These models have been used to improve language understanding across Bing, Office, Dynamics, and other products. Microsoft has made publicly available what is believed to be the largest publicly available AI language model in the world: The Turning model for natural language generation.
Learning from people instead of data at Project Bonsai: Microsoft’s Project Bonsai wants to help businesses teach and manage their autonomous machines. Microsoft notes that Project Bonsai is only the first block of a larger vision to help its customers build autonomous systems. The company also stresses the advantages of machine teaching over other machine learning approaches, especially the fact that it’s less of a black box approach than other methods, which makes it easier for developers and engineers to debug systems that don’t work as expected. In addition to Bonsai, Microsoft also today announced Project Moab, an open-source balancing robot that is meant to help engineers and developers learn the basics of how to build a real-world control system. The idea here is to teach the robot to keep a ball balanced on top of a platform that is held by three arms.
AI evaluates CT findings better than radiologists: Physicians from the Icahn School of Medicine in New York are using two AIs to analyze the CT findings of Covid-19 patients. One of the two softwares uses a neural network to recognize patterns in images. The second AI is intended to be able to make a diagnosis. Both AIs were trained with 534 findings. The test showed that the two AIs achieved better diagnostic results than long-time radiologists.
Virtual conference: “AI practice in businesses” computerwelt.at
Astrophysics: Artificial intelligence discovers gravitational lenses spektrum.de
Facebook: New language AI sounds like a human and is quickly trained mixed.de
Smart Data Analytics: BMW relies on AI in the paint shop automotiveit.eu
Chat: AI to track down fraudsters on Facebook Messenger kurier.at
NUMBER OF THE WEEK
According to expert Dr. Kay Knoche, 80 percent of digital inquiries to companies could be answered by AI-based decision systems.
About the boundaries between humans and AI: The German military is currently debating about autonomous armed drones. Supporters of weapon systems always emphasize that humans have the ultimate decision-making power. “We have no intention of introducing systems that kill autonomously,” says Gerald Funke, head of the Future Development subdivision in the Bundeswehr. Complete AI control could be conceivable in the case of non-armed systems, for example robots that salvage the wounded. A distinction is made between “in the loop”, in which humans are the last resort, and “on the loop”, in which the machine is granted significantly more autonomy. Technology philosopher Catrin Misselhorn points out that people are already excluded from the decision-making process of many programs and still trust them regardless. In many situations, the decision-making processes between man and machine are so interwoven that they can no longer be separated. Researcher Louise Amoore gives an example: “One of the most striking moments in my most recent research was when an experienced surgeon told me that using a surgical robot to cut out tumors changed his assessment of the limits of his own abilities.” The surgeon no longer differentiated between the machine, the algorithms, and his own abilities. Amoore warns: “The same group of algorithms is used in autonomous weapons and autonomous vehicles.”
Geutebrück’s change to an AI service provider: Security company Geutebrück supplies many well-known customers, among them the German Chancellery or the Russian central bank. The company is increasingly focusing on artificial intelligence. The technology is especially helpful during the pandemic. For example, the company develops software that uses video recognition to ensure that people are wearing protective masks. It can also be used to check whether workers on construction sites are wearing their helmets. “Well-known companies from the fields of logistics and industry” have shown interest in the software, said boss Katharina Geutebrück. In 1997, she took over the management of marketing at Geutebrück and has been a member of the family group’s management since 1999. In 2012, she took control of the company. As the hardware market – the company’s specialty – stagnated, the company was increasingly converted into a software manufacturer.
PROJECT OF THE WEEK
Nvidia’s AI builds Pacman clone: Researchers from Nvidia have achieved a functioning AI that can replicate Pac-Man with nothing more than pixels and key presses. The generative adversarial network (GAN) outlined in the research paper, nicknamed GameGAN, is capable of taking pixel and input data from a videogame and recreating a like-for-like carbon copy. It does so without an underlying engine—the AI actually generates a new frame for every on-screen event based on those before it, player action, and a hint of environmental randomness.
“What we can do right now is to achieve impressive results in very narrow domains with machine learning. For example, there are algorithms that can play “Go” superhumanly well. But the same algorithm cannot drive a car, and it cannot learn how the system is designed. At least it is not well suited for this. It is a large construction site and we still have a lot to work on.”
Prof. Dr. Tobias Glasmachers from the Institute for Neuroinformatics at the Ruhr University Bochum on the challenges of modern AI.
AI in the call center: Most call center employees cannot get to their workplaces because of the current pandemic. Artificial intelligence like the Watson Assistant for Citizens from IBM make it possible to answer calls anyway. The pandemic is accelerating a development that has been in the making for a long time: The technological development of voice-based chatbots could soon make human employees in call centers a thing of the past. However, it could take some time before humans are completely replaced by AI in this area.