YouTuber Creates Viral AI Movie Trailer Blending ChatGPT, Midjourney, and More
See How One Creator Leveraged Leading AI Tools To Forge The Next Viral Sensation
Non-invasive speech decoding system developed by Meta AI
- Meta AI researchers develop non-invasive speech decoding method using AI & magnetoencephalography (MEG) brain scans
- Initial study with 175 people shows decoding accuracy around 41% average, up to 80%
- Unlike past invasive methods, doesn't require risky brain surgery
- Major potential to help paralyzed patients and those with speech impairments
- Next step is advancing from decoding perceived speech to producing speech
Slight wage impact but job gains seen from AI in Europe: ECB
- European Central Bank study examines AI's impact on jobs and wages
- Between 2011-2019, AI-exposed occupations saw no change in employment
- But link found between AI automation and increased high-skill jobs, mostly benefiting younger workers
- Contrasts with past automation reducing medium-skill roles
- Results suggest AI currently more likely to complement vs. substitute roles
- Neutral-to-slightly negative impact seen on wages
- Overall effect still uncertain; varies between countries
Creating an AI-powered movie trailer from start to finish
- YouTuber maps process of making own AI movie trailer
- ChatGPT writes 15 scene outline; provides scene prompts and narration guide
- Midjourney & DALL-E 2 used to generate 15 still images
- Runway ML animates images into scenes using motion brush
- ElevenLabs text-to-speech used for narration
- AIVA AI music generator for background score
- Edited together in Davinci Resolve into full trailer
- Shows emerging creative potential blending AI tools across discipline
Creating an AI-powered movie trailer from start to finish
YouTuber Matt Wolfe set out to create his own AI-generated movie trailer from scratch, documenting the full process in a recent video. He started by having the AI chatbot ChatGPT write a 15-scene storyline outline involving a character resembling Leonardo DiCaprio saving his family from an AI apocalypse.
ChatGPT provided visual prompts for each scene as well as suggestions for the trailer narration voice. Matt then generated images for each scene prompt using Midjourney and DALL-E 2, selecting the best outputs and saving 15 final still images.
To animate the still images into a full trailer, Matt used RunwayML's AI video generation capabilities. He mainly relied on Runway's "motion brush" to selectively animate parts of each scene, like adding moving clouds or flaming debris.
For the trailer narration audio, Matt fed ChatGPT's suggested script into ElevenLabs text-to-speech to generate an ominous movie trailer voiceover. He also created suitable background music with AIVA's AI composition tool.
Finally, Matt edited all the pieces together in Davinci Resolve - the 15 animated scene clips, narration audio, and background music - to produce the full creative trailer.
While intended more for entertainment than a commercial product, Matt's experiment shows how AI tools can now collaborate across disciplines - writing scripts, generating images and video, synthesizing audio - opening new possibilities for automating creative media production. The output still requires human judgment, creativity and editing, but AI is taking on more of the rote workload.
6 OF THE BEST AI TOOLS
HEADLIME IS THE GO-TO GPT-3 TOOL FOR MARKETERS.
WRITESONIC IS ONE OF THE BEST ARTIFICIAL INTELLIGENCE-POWERED COPYWRITING GPT-3 TOOLS.
Flywheel launches medical imaging AI platform on Microsoft Azure
US-based medical imaging AI developer Flywheel has launched a new software-as-a-service data management platform on Microsoft Azure. Integrated with NVIDIA's MONAI, the tool aims to accelerate healthcare research and AI model development by facilitating medical imaging data preparation, training, analysis and mobilization. Flywheel is collaborating with Microsoft and NVIDIA to create a robust unified technology platform to empower researchers, clinicians and scientists to enhance innovation, discovery and patient outcomes. According to Trent Norris, Flywheel's Chief Revenue Officer, the partnership brings together proven scalability and AI-enabling solutions to drive global health advances.
Non-invasive speech decoding system developed by Meta AI
Researchers at Meta AI have developed a promising new non-invasive method to decode speech directly from brain activity recordings. The system uses magnetoencephalography (MEG) imaging combined with a dual-module AI model to analyze brain signals and predict the speech heard by participants.
In an initial study with 175 participants who listened to audiobook narrations and sentences, the model achieved decoding accuracies around 41% on average and up to 80% for some participants when analyzing 3 seconds of MEG scans. The decoded phrases were often semantically similar to the target even when incorrect.
Unlike past invasive decoding methods requiring implanted electrodes, this non-invasive approach has major potential for assisting paralyzed patients or those with speech impairments. The next steps are improving the model and advancing from decoding perceived speech to producing speech representations directly from the brain.
While still early and requiring refinement before clinical implementation, the researchers hope this AI-driven MEG imaging technique could one day help patients with limited communication abilities, without needing high-risk brain surgery. The technology also furthers understanding similarities between the brain and AI systems.
Slight wage impact but job gains seen from AI in Europe: ECB
Research by the European Central Bank suggests AI and automation have had a neutral to slightly negative impact on wages but a positive effect on jobs across Europe over the past decade.
Analyzing 16 countries from 2011-2019, occupations with higher AI exposure saw no significant change in employment levels. However, researchers found a link between AI-enabled automation technologies and increased employment, mostly benefiting high-skill jobs and younger workers.
This contrasts with previous automation waves that reduced medium-skill jobs. The researchers see their results as showing AI's potential to complement rather than substitute roles needing advanced expertise - at least for now.
Wage growth for AI-exposed occupations was muted or slightly lower. But the ultimate impact of AI on employment, wages and inequality remains uncertain. Regulatory and economic differences across countries have also led to varying results.
As companies worldwide race to develop new generative AI tools, policymakers are still determining how to govern the technology. With the societal effects still unfolding, striking the right balance between innovation and responsible AI will be key.
Unleash Your Creativity with These 10+ Amazing Free AI Art, Music, and Video Tools
Transform Photos and Videos into 3D Scenes, Generate Original Music, and More with Cutting-Edge AIÂ