AI: from technical challenges to ethical issues. Weekly digest AI: from technical challenges to ethical issues. Weekly digest

New brands and new technologies in the AI space enter the market every day. Asian manufacturers actively join the tech race. Recently, the Chinese e-commerce giant Alibaba announced the release of AI-powered chips. Meanwhile, the Western companies show concern about not only commercial but also an ethical aspect of AI introduction. What is the future of the human workforce if replaced by neural networks? Read the details in the weekly digest by the AI Conference Kyiv press service.

Techstars Berlin to invest $120 000 in Ukrainian startup

Softcube, the Ukrainian startup founded by Oleg Lesov and Oleg Boroday in 2010, was listed in the Techstars accelerator in Berlin. This is the only Ukrainian-based company that was shortlisted during the autumn contest organized by the accelerator. The team will undergo training at Techstars Berlin within October 10-December 6. The accelerator will as well allocate $120 000 to the startup. $100 000 will be invested in the project while $20 000 will cover accommodation costs in Germany.

AI: from technical challenges to ethical issues. Weekly digest - 1

The primary focus of Softcube was Big Data and development of solutions for the online store. Later, the company was substantially changed into the service of video ads powered by AI.

New method of object recognition in photo

A new machine learning method was put forward at the Massachusetts Institute of Technology. Now, AI can more effectively recognize the objects in the images described by the human voice. The technology will be beneficial during the process of development and improvement of voice assistants.

The new technology is based on the convolutional neural network (CNN) consisting of two parts. At the first training stage, one network part was analyzing images while the second was learning a speech spectrogram, or its visual representation. The results of each stage were compared with the right answer. Later, both parts of the algorithm were improved: testers showed the images with several objects and long descriptions to the neural network. AI had to sort the images within the net and single out objects as well find given words in the speech. After that, the neural network compared images to descriptions independently. Watch the video to check out how it works.

As reported by the creators, this technology would allow to improve voice assistants. The modern programs such as Siri and Alexa are based on another more complicated but less effective technology.

SAP’s guiding principles combat massive dismissal in the AI era

The Germany-based IT company SAP has published the ethical guiding principles for AI. The document expresses the company’s vision of the range of social problems and will regulate the development processes in the AI realm. The core idea of the guiding principles lies in the prevention of human workforce discrimination and dismissal.

According to the document, the company is obliged to avoid discrimination of its employees and clients as well as protect their safety and personal data. SAP is also determined to do its best in order to train and retrain students and employees so that they can find the rightful place in the fast-changing world.

SAP is not the first to consider ethical problems at the threshold of the AI era. For example, in September 2016, a lot of such heavyweights as Amazon, DeepMind, Facebook, IBM, and Microsoft established the Partnership on Artificial Intelligence aimed at discussing and solving related issues.

Alibaba works on a quantum processor and AI chips

AI: from technical challenges to ethical issues. Weekly digest - 2

The commercial giant Alibaba develops quantum solutions. Recently, the company announced the release of semiconductors. The new department will work on brand-new quantum PCs, quantum chips, and AI-powered processors. The members of DAMO Academy, a research department of Alibaba, and the CPU manufacturer C-SKY Microsystems (bought by Alibaba in April) will be the core of the developing team.

As reported by the Chief Technology Officer at Alibaba Zhang Jianfeng at the annual company’s conference in Hangzhou, the first AI-fueled chip was called AliNPU and would be presented in Q3-4 2019.

Moreover, the company plans to introduce quantum processors in 2-3 years. As explained by Zhang Jianfeng, the principal target of the new production is to make Alibaba independent from the U.S. import: when the international relations become tough, this appears to be a speculation topic.

Join us at AI Conference Kyiv to take part in the discussion of the AI future in Ukraine. Early bird tickets are available on the website.

Related news