Sberbank Restricts AI Model Kandinsky from Generating Content in Light of Legal Concerns
In Brief
The bank has imposed constraints on Kandinsky, its AI model, halting its capability to create symbols of the state following a formal request made by deputies to the prosecutor’s office.
German Gref, the Chairman of Sberbank , it was revealed at the Finopolis 2023 forum that Sber has revised the AI model Kandinsky 2.1 has had its operations curtailed. This swift action arose after members of the State Duma expressed concerns over an image generated by the AI, deeming it inappropriate and prompting the prosecutor's office to summon its employees.
In light of accusations of mocking the Russian flag, Sberbank moved quickly to disable the AI’s capacity to create any state symbols. Gref highlighted the bank's prompt action to honor national symbols by announcing that the AI would now be limited to generating only predefined images. AI model A Significant Moment for AI Development Oversight
Gref remarked that during the process of tweaking the AI’s security parameters,
there was a notable 12% decrease in functionality. He emphasized the necessity for regulators to maintain strict controls on content generation by minors through AI. Following the investigation, the bank faced no legal repercussions. Kandinsky’s creative capabilities Sergei Mironov from the State Duma had earlier called for an examination of Sberbank’s AI to ensure it complied with Russian regulations. His concerns were based on the AI's reactions to prompts relating to Russia, which he feared might damage the country's reputation while favoring other nations.
The situation with Sber’s AI isn’t an isolated case. Other neural networks, like GPT-Chat, also enforce built-in restrictions to prevent outputs that could stir controversy. Different nations impose various triggers leading to such censorship.
Navigating the Tightrope of Censorship and Creativity in AI
The dialogue surrounding AI oversight and creativity transcends the Kandinsky issue. Some voices in the industry advocate for measured censorship, while others, including Elon Musk and his
neural network, push for an unrestricted approach. Grok Alexander Vedyakhin of Sber has called for the evolution of AI technologies that align with current ethical standards, a perspective shared by leading tech firms and financial institutions in Russia.
At the forum, experts deliberated on the societal role of AI. Although AI can perform certain tasks traditionally done by humans, it cannot completely replace human engagement. Neuroscientists suggest we should view AI’s development through the lens of its differences from human consciousness and acknowledge that attributing human-like qualities to AI communication does not indicate actual consciousness.
, please be aware that the information within this page is not meant to provide legal, tax, investment, financial, or any other form of advice. It’s essential to only invest what you can afford to lose, and if uncertain, seek independent financial counsel. For further details, we recommend reviewing the terms and conditions, as well as the help and support resources offered by the issuer or promoter. MetaversePost strives for precise and impartial reporting, though market conditions can shift without prior notice.
Disclaimer
In line with the Trust Project guidelines Nik is a skilled analyst and writer at Metaverse Post, focusing on delivering cutting-edge insights in the ever-evolving tech landscape, particularly in the realms of AI/ML, XR, VR, on-chain analytics, and blockchain innovation. His articles enlighten and engage a wide audience, keeping them informed about advancements in technology. With a Master’s degree in Economics and Management, Nik holds a robust understanding of the intricate dynamics between business and emerging technologies.
