- AI Trends
- Posts
- π€ AI Reveals Herculaneum Scrolls ποΈ
π€ AI Reveals Herculaneum Scrolls ποΈ
Good morning! Here's your AI Trends dose, acting as your favorite author, spinning a captivating narrative of mystery, intrigue and wondrous revelations in the enthralling plotline of AI, tech, and innovation.
Let's flip open our book and delve into today's exciting news chapters:
1. Artificial Intelligence Reveals Secrets of Herculaneum Scrolls ποΈπ€
2. How Much Do Artists Really Earn from Generative AI? π€
3. DeepMind and Trust: A Debate on Ethics in AI π§π
In this edition, we journey back in time to the ancient city of Herculaneum, whose secrets are being unraveled not by a daring archaeologist with a hat and whip, but by cutting-edge artificial intelligence. This technological achievement promises to unlock literary mysteries from nearly two millennia ago.
Luke Farritor, a brilliant mind from the University of Nebraska-Lincoln, brewed an algorithm mixed with AI that managed to decode a word from the charred Herculaneum scrolls. The Roman city was buried following the eruption of Vesuvius in 79 A.D., and among its ruins, a cache of 1,800 scrolls was found which could potentially contain valuable historical literary and philosophical works. But they've always been a crackling enigma, too fragile to manually unroll.
π» This would be the modern equivalent of deciphering a forgotten cryptocurrency wallet password, but instead of bitcoins they are valuable literary works from ancient Rome!
[ Made with AI. Instructions: AI decoding secrets ]
This milestone is impressive on all levels and opens a new path in the field of digital archaeology. Farritor's AI algorithm could be the chosen method for future decoding of the remaining scrolls, retrieving a glorious library lost in pre-digital times.
The Vesuvius Challenge represents a successful crossroads of technology and humanities, where AI aids in uncovering our far-off past in ways inconceivable merely a decade ago. With researchers from around the world vying for the grand prize of the Challenge, we can be assured that more phenomenal discoveries are on the horizon.
With the rise and monetization of generative AI, creators whose works are used to train these technologies are demanding their fair share of the pie. Prominent and emerging tech companies are in the spotlight as artists, authors, and musicians seek fair compensation for the use of their works in the development of AI models.
A group of creators, backed by more than 8,500 writers, including notable names like Margaret Atwood, Dan Brown, and Jodi Picoult, are raising their voices against the unauthorized use of their works. Companies like Adobe and Getty Images are introducing compensation systems, although the lack of transparency and specificity in the details of these systems is creating an atmosphere of uncertainty and frustration among the creative community.
π¨ It seems that artists might need generative AI to decode these companies' compensation policies.
[ Made with AI. Instructions: Artists striking, cartoon ]
The pursuit of equitable and transparent solutions is imperative. It is crucial that companies develop clear and fair compensation models that appropriately value the creators' contributions.
Only then can a balance be achieved in which technological innovation and artistic creativity coexist and thrive together, paving the way towards a more collaborative and respectful future of creators' rights.
This week, Google-owned artificial intelligence research and development lab DeepMind published a paper proposing a framework for evaluating the ethical and social risks of AI systems. This move could be a crucial step forward in maintaining ethics in our growing relationship with Artificial Intelligence, but can we trust DeepMind to be ethical?
The report proposes different degrees of involvement from AI developers, application developers, and public "stakeholders" in the evaluation and auditing of AI systems. Coincidentally, the proposal comes just before the AI Safety Summit, a United Kingdom government-sponsored event that will bring together international governments, leading AI companies, civil society groups, and research experts to focus on how to manage the risks of recent advancements in AI.
π€ Is DeepMind trying to be ethical or is it just playing the good kid on the AI playground?
[Made with AI. Instructions: Aristotle as an AI robot]
In analyzing this proposal, it is interesting to take a look at how DeepMind's parent company, Google, stacks up in a recent study conducted by Stanford researchers that ranks the ten main AI models in terms of how openly they operate. According to 100 criteria, PaLM 2, one of Google's flagship text analysis models, scores a scant 40%. Even though DeepMind did not directly develop PaLM 2, the fact that its parent company falls short on key transparency measures suggests there is not much pressure from the top for DeepMind to do a better job.
On the contrary, DeepMind seems to be taking steps to change the perception of being secretive about its models' architectures and inner workings. A few months ago, the lab, along with OpenAI and Anthropic, committed to providing the UK government "early or priority access" to their AI models to support research on evaluation and safety.
AI Bites πͺ
Appleβs job listings suggest it plans to infuse AI in multiple products. Apple seems to be finally getting serious about infusing generative AI into its products after announcing a solitary βTransformerβ model-based autocorrect feature on iOS 17 earlier this year.
Google Pixelβs face-altering photo tool sparks AI manipulation debate. The camera never lies. Except, of course, it does - and seemingly more often with each passing day.
Smart upcycling machine dissects batteries to save them. Circu Li-ion cofounder and CEO Antoine Welter says his startupβs upcycling machine can diagnose batteries in seconds.
Watch this, thank me later... π
π Adobe has launched a digital dress, known as "Project Primrose". This dress is part of an initiative to apply designs created with Adobe apps in the physical world, allowing users to send their created patterns to this digital dress.
The dress can magically change color and pattern thanks to the incorporated technology. It was introduced during Adobe's annual event, Adobe MAX 2023, and is covered with sequins that utilize liquid crystals to create miniature screens π
How Did You Find Today's Edition?Your feedback is invaluable to us. Help us refine our content by letting us know how you felt about today's edition. |