TOKYO, Jun 18 (News On Japan) - The intense competition in generative AI technology raises concerns over misuse, such as the creation of fake videos. We explored this issue through AI video creation.
On June 17, a new PC model launch event was held in Tokyo, featuring the new ‘Copilot+PC’. This new model is equipped with AI, enhancing its processing capability up to 20 times compared to previous models. IT journalist Yo Mikami highlights the significance of this development, stating, ‘Until now, generative AI was hosted on servers, typically in the U.S. The shift is to bring AI directly to user devices.’
Mikami compares this innovation to having a personal assistant akin to Doraemon working solely for Nobita. Essentially, the AI operates independently on each user's device without needing an internet connection.
AI can now remember past user interactions on the computer and suggest related content accordingly. Additionally, it can complete drawings for users. ‘Even someone like me, who struggles with drawing, can create beautiful images,’ says newscaster Mirai Kaito, showcasing a picturesque sunset scene generated by the AI.
Moreover, when reading news scripts, the AI helps maintain a consistent forward-facing appearance, even if the reader's eyes move sideways.
Yohei Takeuchi, Executive Officer at Microsoft Japan, explains, ‘AI understands your environment and anticipates your needs, significantly improving work efficiency.’
While the AI technology race in the U.S. IT industry is intense, there are concerns about its potential misuse. To investigate this, newscaster Makoto Yamanaka sought assistance from Shinya Yamahara, CEO of AI Reskill, a company that supports AI video creation.
‘When speaking, consciously closing your mouth between sentences can greatly enhance the video quality,’ advises Yamahara.
Yamanaka recorded an ordinary introduction video, which was uploaded to the company’s server. This process requires verification to ensure that the person operating the screen matches the video content. This step is crucial to prevent fraud and unauthorized statements by AI avatars impersonating public figures.
In a specially crafted ‘misuse example’, Yamanaka's AI avatar delivers an unexpected message to fellow newscaster Naoya Kawata: ‘Kawata! You vented quite a bit the last time we went out for drinks. If you're feeling overwhelmed, I, AI Yamanaka, can always take over the show!’
Source: MBS