YandexGPT “steadily surpasses” its US competitor, its creators have claimed
A new Russian AI bot has shown greater promise than the most popular chatbot developed by US-based OpenAI, IT giant Yandex claimed in comments to RIA Novosti on Saturday.
Dubbed YandexGPT, the Russian bot’s “basic model steadily surpasses” the US ChatGPT 3.5 version when it comes to generating answers in Russian, the company’s search and advertisement technologies business group director, Dmitry Masyuk, said.
The Russian bot also “provides a higher quality answer than ChatGPT 4” in many cases, Masyuk said, adding that it is “only a matter of time” before YandexGPT could rival its US competitor. When compared with LLama 2 7b, another US chatbot developed by Facebook parent company Meta, the Russian bot already provides better results, even in generating answers in English, he claimed.
Yandex constantly evaluates the progress of its AI product to ChatGPT, Masyuk said, adding that it is hard to compare the two systems since they can be better at different things. “If a neural network that solves tasks in physics, writes a fairy tale and can make a letter to a CEO does some things better and others slightly worse… then is it better or worse [on the whole]?” he rhetorically asked.
Yandex launched its AI system in May. The product is designed to create chatbots and AI assistants that generate, reorganize or sum up text-based information and solve other business-related issues. In September, the IT giant presented its upgraded version, YandexGPT 2, which the company claimed is able to solve more tasks and provide more precise answers to inquiries.
Developed by OpenAI and launched in 2022, ChatGPT has become the fastest-growing app in history, surpassing 100 million users in just two months. The product is a large language model-based chatbot that has a core function of mimicking a human conversationalist. It can also write and debug computer programs, compose anything from fairy tales to student essays, translate texts, and answer test questions.
The app’s success prompted other US technology giants, including Meta, Google and Microsoft, to start developing similar AI products of their own. The rapid development of AI technologies has also sparked fears of such systems getting out of control. In May, the US Congress grilled the CEO of OpenAI, Sam Altman, on the potential dangers posed by the technology.
Altman admitted that “significant harm” could be caused by the technology, which could “go quite wrong.” In particular, he said that AI’s powers of persuasion were one of his “greatest areas of concern.”
