philip lelyveld The world of entertainment technology

... At the heart of this research lies the ARC benchmark, a sophisticated tool that challenges AI systems to demonstrate genuine understanding rather than mere memorization. Unlike traditional benchmarks, which often rely on pattern recognition, the ARC benchmark pushes AI into uncharted territories of thought, requiring it to solve problems creatively and adaptively. Coupled with innovative methods like test-time training, which allows AI to learn and adapt in real-time, these advancements have propelled AI models to achieve human-level reasoning—and even beyond. ...

TL;DR Key Takeaways :

• MIT’s study highlights a novel approach to abstract reasoning as a key to achieving Artificial General Intelligence (AGI), using the ARC benchmark to test machine intelligence creatively and adaptively.
• The ARC benchmark serves as an IQ test for machines, assessing AI’s ability to apply abstract reasoning to novel situations, marking a significant leap in reasoning capabilities.
• Test-time training, which updates model parameters during inference, has enabled AI systems to surpass human-level reasoning on the ARC benchmark, representing a major advancement toward AGI.
• Search algorithms enhance AI’s problem-solving abilities by allowing efficient exploration of solutions, crucial for achieving human-level performance and advancing toward AGI.
• The study indicates AI models have surpassed human reasoning on the ARC benchmark, suggesting AI can now perform tasks once exclusive to human intelligence, paving the way for future AGI development.

See the full article here: https://www.geeky-gadgets.com/artificial-general-intelligence-advancements/

OpenAI to Launch Autonomous AI Agent

OpenAI is set to launch an autonomous AI agent, code-named "Operator," early next year. ...

Unlike traditional AI interfaces (which work on specific tasks), Operator and its counterparts will actively engage with your computer, completing tasks without human intervention. In success, the agents will jump from program to program doing stuff like cutting and pasting, answering emails, doing rudimentary computer maintenance, and more. ...

"Agentic," which is just a fancy word for agency, is a new vocabulary word that will be on this week's test. All kidding aside, agents are the future. We'll see if the pre-canned, purpose-built computer agents are going to live up to the hype. ...

See the full post here: https://shellypalmer.com/blog/

A Wannabe ‘Pixar of AI’ is Signing Up Stars

...

In a nutshell, Asteria is using AI to speed up projects and help human artists make better use of their time by using tech that’s trained on their work to do things like test out different visual concepts or convert storyboards to animatics. 

The ultimate goal of all of this is to become the “Pixar of AI” and, in fact, they’ve hired some employees who were laid off from the storied company. (Not to say that established animation heavyweights like Disney, Pixar and DreamWorks aren’t using AI. They are.)

The tech helps them keep costs down, and Mooser tells me the sweet spot is under $10 million per project. That means they’re able to take risks on projects that wouldn’t see the light of day at a traditional studio — whether because they’re too expensive or not in line with that particular company’s current brand persona — and use an indie financing model that gives backend to creatives. ...

In this issue, you’ll learn . . .

• How a potential new AI-based studio model could soon become the norm
• How much time Asteria’s AI tools promise to shave off development — and the opportunities that creates for an indie studio to compete
• How Asteria is trying to solve AI’s original sin
• Why the company believes movies made with its AI tools will still be copyrightable
• Asteria’s bold profit-participation model — and whether talent dealmakers think it can work
• The major questions top lawyers ask before considering a backend deal
• Why dealmakers say Hollywood’s attitudes about AI are changing

...

See the full story here: https://theankler.com/p/asteria-film-wannabe-pixar-of-ai-natasha-lyonne-bryn-mooser?utm_source=substack&utm_medium=email&utm_content=share

Upcoming Changes to AI Regulation in the New Administration

With the second Trump Administration set to take power in January 2025, one can expect a pendulum swing in many aspects of technology policy. For example, while it is expected that President Trump will will continue efforts by the Biden Administration to limit China’s access to advanced semiconductor technology, the SEC is expected to be much more crypto-friendly. What is unclear, however, is how the new federal government will address AI regulation.

It’s a good bet that a Trump White House will take a lighter approach to the oversight of the AI industry than the preceding Biden Administration and do little to upset the current favorable environment for AI investments and start-up companies. Regardless of the exact route forged by the new Trump Administration, the U.S. already dominates private-sector investment in AI, and the government will likely take efforts to nurture this trend. ...

See the full story here: https://natlawreview.com/article/upcoming-changes-ai-regulation-new-administration

The impact of AI on the film industry

...

Casting and location scouting: AI can streamline the casting process by analyzing facial expressions, voices, and physical characteristics to match actors with suitable roles. Examples of this technology already exist: for instance, Casting Matching Process AI (CMP-AI) creates the character profile by analyzing it directly from the script, based on keywords using three narrative elements: physical-aesthetic, psycho-aptitude and scenic. Also, it can aid location scouting by analyzing hundreds of images or providing insights about weather predictions, crowd control and logistics.

...

As expected, one of the primary concerns is how AI uses copyrighted material as input for training its models. The 2019 Directive on Copyright in the Digital Single Market introduced exceptions for text and data mining (TDM), allowing rights holders to control or exclude their content from these processes. The proposed AI Act builds on this by requiring AI model providers to respect copyright and provide detailed summaries of the data used for training. But a recent landmark judgment issued by a German Court (Kneschke v. LAION), which held that a nonprofit organization copying images didn't infringe copyright law, may have significant impact in the application of the TDM exceptions. For example, according to the ruling, creators cannot opt out of having their work used for AI training in scientific research. ...

The draft law on AI approved in April in Italy proposes to revise national copyright law by clarifying that protected "intellectual works" must be of "human" origin and that AI-assisted works may still qualify for protection if their creation derives from the author's intellectual work. The draft law also aims to complement European legislation by facilitating the identification and recognition of AI systems in the creation of textual, photographic, audiovisual and radio content. ...

See the full story here; https://www.lexology.com/library/detail.aspx?g=ca7cb00c-1668-4838-8f86-c10aeda6e2cb

ChatGPT blocked 250,000 AI image requests of US election candidates

ChatGPT refused more than 250,000 requests to generate images of the US election candidates using their artificial intelligence (AI) platform. 

OpenAI, the company behind the AI chatbot, said in a blog update on Friday that their platform DALL-E, used to generate images and video, rejected requests to make images of president-elect Donald Trump, his choice for vice president JD Vance, current president Joe Biden, democratic candidate Kamala Harris, and her vice-presidential pick, Tim Walz. 

The refusals were due to “safety measures” that OpenAI put in place before election day, the blog post said. 

“These guardrails are especially important in an elections context and are a key part of our broader efforts to prevent our tools being used for deceptive or harmful purposes,” the update read. ...

The company said in August it stopped an Iranian influence campaign called Storm-2035 from generating articles about US politics and posing as conservative and progressive news outlets.  ...

See the full story here: https://www.euronews.com/next/2024/11/11/chatgpt-blocked-250000-ai-image-requests-of-us-election-candidates

Anthropic, Palantir, Amazon team up on defense AI

Palantir and Anthropic are partnering with Amazon Web Services to make Anthropic's Claude models available to U.S. intelligence and defense agencies, the companies announced Thursday.

Why it matters: As Washington rushes to bring AI to every nook and cranny of government, companies that have already mastered the public sector's complex contracting requirements have a leg-up.

The companies said in a release that access to the Claude models from within Palantir's data analytics platform will help agencies with tasks like:

• "processing vast amounts of complex data rapidly, 
• elevating data driven insights, 
• identifying patterns and trends more effectively, 
• streamlining document review and preparation, 
• and helping U.S. officials to make more informed decisions in time-sensitive situations while preserving their decision-making authorities." 

Between the lines: Contractors providing software services to government agencies, particularly in high-security applications, need to be accredited under a variety of classifications.

...

See the full story here: https://www.axios.com/2024/11/08/anthropic-palantir-amazon-claude-defense-ai?utm_source=substack&utm_medium=email

Trump’s Second Presidency Could Mean AI Chaos, or a Flurry of Innovation 

...

Nobody can tell if removing U.S. safeguards addressing AI development would result in locally sourced AI disasters, of course, and it’s likely that some state safeguards would remain in force, with California authorities, for example, very much on top of this idea. But as TechCrunch pointed out on election day, while “all” the AIs the news site looked at were “behaving” and ensuring responsible and correct info was being shared by their systems when the public asked for election info, one AI was significant for its openness to sharing blatant misinformation: Musk’s Grok.

Is this a sign that under the next Trump presidency the public really should fear AI, and its implications for stealing people’s jobs and violating their privacy? ...

See the full story here: https://www.inc.com/kit-eaton/trumps-second-presidency-could-mean-ai-chaos-or-a-flurry-of-innovation/91000751

AI Robot Artist’s Turing Portrait Fetches $1.1 Million at Sotheby’s

A portrait of computer science pioneer Alan Turing by humanoid robot artist Ai-Da sold for $1.08 million at Sotheby's, marking another milestone in AI art's rapid market ascent. The sale price exceeded pre-auction estimates building on momentum from previous AI art breakthroughs. ...

"The combination of art and technology is gaining more interest," he said, "particularly among a younger generation of collectors." ...

But as controversial as it is, many artists are already embracing generative AI. From big names like Refik Anadol’s Machine Hallucinnations to the works of amateur enthusiasts like Terrance Washington’s “Country Woman” the fusion of human creativity and machine learning seems to be more intertwined.

And this is exactly the message Ai-Da wanted to deliver with its artwork: Humans should reflect about the role of AI in the future of human society. ...

And AI bots have proven very capable of making money all by themselves—not just by making art. Most recently, an AI chatbot called Terminal of Truths generated enough social media influence to propel a cryptocurrency, Goatseus Maximus, into the top 100 digital assets by market capitalization, surpassing established tokens like IOTA and Zcash. ...

Major online art forums have banned AI-generated works, with some communities going so far as to exclude human artists whose style resembles AI outputs, so a fully AI-generated artwork being sold by a prestigious auction house is an important endorsement for generative AI advocates. ...

Ai-Da's "AI God" portrait was displayed at the UN's 2024 AI for Good Global Summit. The robot artist spent up to eight hours completing each of the 15 paintings in the collection, using fragmented imagery to comment on society's algorithmic transition. ...

Cuba Elliott, a curator specializing in artificial intelligence in creative industries, noted to Decrypt that while AI art auctions aren't new, Ai-Da's humanoid approach "helps to further broaden the definition of AI art."

See the full story here: https://decrypt.co/290899/ai-robot-artist-turing-portrait-fetches-1million-sothebys

The directors of A24’s new horror film tell BI why they used their movie to slam Hollywood’s ‘weird’ use of AI

...

Woods told BI that he doesn't understand why it's legal for generative AI to take in existing creations and use them to regurgitate new work, profiting the company that owns the AI without compensating the original creators.

"Stealing other people's things and then generating something that's better in a matter of seconds is insane," Woods said. ...

Most recently, the horror studio Blumhouse Productions caught flak for partnering with Meta on a new initiative that will allow select filmmakers, including Casey Affleck, to test out the tech company's generative AI video system, Meta Movie Gen.

The system uses text inputs to create and edit realistic videos; in this partnership, the filmmakers would use the AI-generated video clips within larger works. ...

See the full story here: https://www.yahoo.com/entertainment/directors-a24s-horror-film-used-093102076.html