Wednesday 28 June 2023

Underpowered Artificial Intelligence: Russias Latest Weapon

The Kremlin’s Plan to Control Information Through AI Faces an Insurmountable Flaw: Facts

In a stunning display of AI manipulation, an ethereal, electric blue video of Russian far-right leader Vladimir Zhirinovsky declared at the St. Petersburg International Economic Forum that “Ukraine will be liberated from the Nazis.” However, Zhirinovsky died last year, and his confident declaration clashes with the reality of Ukraine’s situation. The Kremlin’s desire to use AI to increase control over information is hindered by the presence of facts.

The St. Petersburg International Economic Forum: A Diminished Western Presence

The St. Petersburg Forum, an annual gathering of tech industry leaders, was intended to attract Western investment. However, this year, no Western leader attended, and the guest list was dominated by individuals from India, China, and Arab countries. Despite this diminished Western presence, Russian President Vladimir Putin used his keynote speech to highlight Russia’s progress in AI, specifically automated trucks and self-driving taxis. However, his main concern was the potential dominance of Western-directed AI and the need to protect Russia’s national security and citizen interests.

The Danger of AI Assumptions: Language Models and Propaganda

Putin’s concern about Western-dominated AI is not unfounded. Large language models like Chat GPT rely on patterns found in training data, predominantly in English. As a result, when a Russian-language prompt was submitted to ChatGPT asking about color revolutions, the chatbot responded with an interpretation that contradicted the official Russian media’s narrative. This discrepancy highlights the inherent limitations and biases of AI models trained on specific data sources.

Russia’s Ambitions to Rival ChatGPT: Yandex, Sistemma, and Sberbank

Despite the potential limitations and biases of Russian-trained models, several major Russian companies, including Yandex, Sistemma, and Sberbank, have announced their ambitions to rival ChatGPT. Yandex has incorporated AI into its ‘Alice’ offering, Sistemma unveiled its own ChatGPT competitor based on research from Stanford, and Sberbank debuted a beta version of ‘GigaChat’ with image generation functionality. However, these Russian models face strong headwinds due to the supremacy of US versions and Western tech sanctions.

Challenges Faced by Russian AI Offerings: Computing Power and IT Talent

Intense computing power is essential for the most powerful AI models, and this power is scarce in Russia due to Western tech sanctions. Retrieving results from Russian models may be significantly slower than accessing American alternatives, even with the use of VPNs. Additionally, the invasion of Ukraine led to a mass flight of IT talent, limiting the brainpower available for AI development in Russia. These challenges reinforce the fundamental conflict between the Kremlin’s desire for information control and the open-ended possibilities of generative AI.

Putin’s Dilemma: AI Potential versus “Information Security”

Ultimately, Putin must decide whether to prioritize the potential of AI or maintain his cherished “information security.” However, it appears that his quest for information control is more likely to fail in the face of an evolving technological landscape. The nationalistic ideologies once championed by figures like Vladimir Zhirinovsky may be fading into oblivion as AI continues to reshape the information landscape.

Editor Notes: GPT News Room

In conclusion, the Kremlin’s plan to use AI as a means of controlling information is hindered by the presence of facts and the limitations of AI models. Despite the ambitions of major Russian companies, the dominance of Western versions and tech sanctions pose significant challenges. Furthermore, the mass flight of IT talent from Russia following the invasion of Ukraine limits the country’s AI capabilities. It remains to be seen whether Putin will prioritize information control over the potential of generative AI. For more news and analysis on AI and other tech-related topics, visit the GPT News Room.

Source link



from GPT News Room https://ift.tt/faZpiu3

No comments:

Post a Comment

語言AI模型自稱為中國國籍,中研院成立風險研究小組對其進行審查【熱門話題】-20231012

Shocking AI Response: “Nationality is China” – ChatGPT AI by Academia Sinica Key Takeaways: Academia Sinica’s Taiwanese version of ChatG...