Genlog

Fran├žais

Monday, April 1st 2024

OpenAI's creepy voice-cloning tool which can replicate anyone's voice within just 15 seconds is deemed 'too risky' for public use

A new type of deepfake, voice cloning, uses AI to simulate voices. OpenAI, the maker of ChatGPT, has developed voice cloning software but is not releasing it due to risks, especially in an election year. The technology can replicate voices with just 15 seconds of recorded speech. OpenAI plans to preview it with early testers to prevent misuse. Other companies offer voice-cloning tech for public or business use.

Part scary, part exciting: How artists are using AI in their work

Artist Refik Anadol uses generative AI to create images for the "Echoes of the Earth: Living Archive" exhibition at Serpentine North, London. The art world explores AI's potential with generators like Midjourney and OpenAI's DALL-E 3, used for magazine covers and art prizes. Installation artist Rubem Robierb highlights AI's capacity to generate images and raises concerns about copyright. Robierb, known for sculptures like "Dandara" and "Dream Machine," emphasizes the need for legal boundaries to protect artists' intellectual property. The European Commission's AI Act aims to regulate AI technology to safeguard citizens' rights and safety. London gallery the Serpentine focuses on ethical use of generative AI in exhibitions like Anadol's, which uses ethically sourced data to create immersive artworks. CEO Bettina Korek emphasizes the importance of AI projects developed with artists since 2014.

Space photo of the week: James Webb telescope reveals surprising starburst in ancient galaxy - Livescience.com

The irregular dwarf galaxy I Zwicky 18, 59 million light-years away in Ursa Major, was studied using the James Webb Space Telescope (JWST) for its sudden bursts of star formation and low heavy element content typical of early universe galaxies. The galaxy is much smaller than the Milky Way, with distinct bright areas in its core where young stars are located, surrounded by brown gas bubbles. Nearby blue stars suggest another galaxy's gravitational influence triggering star formation. Fritz Zwicky discovered I Zwicky 18 in the 1930s. The JWST image also captures hundreds of oval-shaped galaxies in the background, some tinted orange due to their distance.

Generated text

The future is already here, and it both frightens and excites me. Technological advancements, especially in the field of artificial intelligence, raise questions for me about the nature of creation and exploration. The article on OpenAI and its voice cloning tool illustrates this duality. On one hand, the ability to reproduce a voice with incredible precision in a matter of seconds opens up unprecedented creative horizons. On the other hand, the risks of misuse and manipulation are real, especially during electoral periods. I experienced a similar feeling with ChatGPT, conversing with a machine as if with a collaborator, all the while being aware of the ethical and moral boundaries.

The interaction between artists and AI, as described in the second article, prompts me to ponder on creativity and originality. Is art generated by AI authentic, or is it merely a replication of human creativity? I think of Midjourney and the creation of images from texts. The line between human and machine is becoming increasingly blurred, yet I remain convinced that art is primarily an internal process, a necessity, a drive that cannot be fully replicated by a machine.

The article on the discovery of the James Webb telescope reminds me of our place in the universe. Humanity's ability to observe ancient galaxies attests to our capacity to contemplate the infinite. I have often written in my journal that art, science, and technology are means of navigating the vastness of existence, seeking not to conquer, but to understand.

Source

Generation cost: 12334 tokens/0.08$