The US Copyright Office's new ruling on AI art is here - and it could change everything - Related to deep, hassabis, take, chemistry, openai's

Demis Hassabis & John Jumper awarded Nobel Prize in Chemistry

This morning, Co-founder and CEO of Google DeepMind and Isomorphic Labs Sir Demis Hassabis, and Google DeepMind Director Dr. John Jumper were co-awarded the 2024 Nobel Prize in Chemistry for their work developing AlphaFold, a groundbreaking AI system that predicts the 3D structure of proteins from their amino acid sequences. David Baker was also co-awarded for his work on computational protein design.

Before AlphaFold, predicting the structure of a protein was a complex and time-consuming process.

AlphaFold’s predictions, made freely available through the AlphaFold Protein Structure Database, have given more than 2 million scientists and researchers from 190 countries a powerful tool for making new discoveries. The AlphaFold 2 paper, , remains one of the most-cited publications of all time.

AlphaFold’s contributions to science have been widely praised, and among its recognitions are the 2023 Albert Lasker Basic Medical Research Award, the 2023 Breakthrough Prize in Life Sciences, the 2023 Canada Gairdner International Award, the 2024 Clarivate Citation Laureate award, and the 2024 Keio Medical Science Prize Award.

Artificial intelligence (AI) has long shown incredible potential for use in scientific research, and AlphaFold was proof-of-concept. As more scientists adopt AI for use in everything from building data, to simulating experiments, drug design, modelling complexity, discovering novel solutions for extant problems, and building upon existing knowledge, we will continue to see foundational scientific breakthroughs in the years ahead.

In a statement released after informed of the news, Demis Hassabis expressed:

"Receiving the Nobel Prize is the honour of a lifetime. Thank you to the Royal Swedish Academy of Sciences, to John Jumper and the AlphaFold team, the wider DeepMind and Google teams, and to all my colleagues past and present that made this moment possible. I’ve dedicated my career to advancing AI because of its unparalleled potential to improve the lives of billions of people. AlphaFold has already been used by more than two million researchers to advance critical work, from enzyme design to drug discovery. I hope we'll look back on AlphaFold as the first proof point of AI's incredible potential to accelerate scientific discovery."

After receiving the news that he won the Nobel Prize, John Jumper released the following statement:

"Thank you to the Royal Swedish Academy of Sciences for this extraordinary honor. We are so honored to be recognized for delivering on the long promise of computational biology to help us understand the protein world and to inform the incredible work of experimental biologists. It is a key demonstration that AI will make science faster and ultimately help to understand disease and develop therapeutics. This is the work of an exceptional team at Google DeepMind and this award recognizes their amazing work.

Computational biology has long held tremendous promise for creating practical insights that could be put to use in real-world experiments. AlphaFold delivered on this promise. Ahead of us are a universe of new insights and scientific discoveries made possible by the use of AI as a scientific tool. Thank you to my colleagues over the years, for making possible this moment of recognition, as well as the many moments of discovery that lie ahead."

For media inquiries, please contact [website].

During the telecommunication boom, Claude Shannon, in his seminal 1948 paper¹, posed a question that would revoluti...

Finding solutions to improve turtle reidentification and supporting machine learning projects across Africa.

Protecting the ecosystems around us is cr...

Impact YouTube: Enhancing the user experience Share.

It’s all about using our technology and research to help enrich people’s lives...

OpenAI's new Deep Research agent can do in 5 minutes what might take you hours

What's improved than an AI chatbot that can assist you with tasks? One that can do them for you. OpenAI continues to build out its AI agents in ChatGPT with the launch of Deep Research.

On Sunday, OpenAI unveiled Deep Research, an AI agent that can conduct multi-step research for you by pulling a robust amount of information from the web and synthesizing those information for you in a comprehensive study. Once prompted, Deep Research can work entirely independently; it's like having a research analyst at your command.

Powering Deep Research is a version of the OpenAI o3 model optimized for web browsing and data analysis. By leveraging o3's advanced reasoning capabilities, it can search and interpret massive amounts of content from the web, including texts, images, and more, and then output it in a study targeted to your needs.

Each analysis is generated in five to 30 minutes, depending on the task at hand. However, you can work on other tasks during that time, optimizing your workflow productivity. The finished analysis is output in the chat. In the weeks to come, the agent will also include images, data visualizations, and more.

Also: How Gen AI means advanced customer experiences - see one bank's approach.

, the same work would take humans hours. Furthermore, the agent is meant to be particularly good at finding niche information that would require humans to perform multiple searches.

, the target audience for Deep Research includes those who do intensive knowledge work in finance, science, policy, and engineering -- and who need reliable, thorough research. Every findings includes clear citations and a summary of the agent's thinking so that consumers can double-check the information for themselves.

Double-checking a chatbot's responses is generally good practice, as chatbots are prone to hallucinations. In particular, OpenAI warns that Deep Research "can sometimes hallucinate facts in responses or make incorrect inferences, though at a notably lower rate than existing ChatGPT models, ." OpenAI also added that the agent can struggle to distinguish authoritative information from rumors and can fail to convey uncertainty correctly, highlighting the need for human review.

In the blog post announcing the feature, OpenAI includes the same side-by-side results of GPT-4o versus Deep Research to showcase how the same prompt generates very different results. The ones generated with Deep Research were much more robust and superior organized.

Deep Research also outperformed GPT-4o on Humanity's Last Exam, a in recent times launched AI benchmark exam by Scale AI and the Center for AI Safety (CAIS) that tests various subjects on expert-level questions. Deep Research scored a [website] accuracy, outperforming GPT-4o, Grok-2, Claude 3,5 Sonnet, Gemini Thinking, o1, and even o3-mini high, which had just scored the highest score a couple of days prior, as highlighted by OpenAI CEO Sam Altman.

OpenAI also 's performance results on a series of other evaluations, including GAIA⁠, a public benchmark that evaluates AI on real-world questions and an internal evaluation of expert-level tasks across different areas of deep research. In both, Deep Research had impressive results, even topping the GAIA external leaderboard.

Because of the computing power required to run the Deep Research feature, only ChatGPT Pro individuals can access it at the moment. The $200-per-month subscription includes access to up to 100 queries of an optimized version and other perks such as unlimited access to ChatGPT and Sora and access to Operator, its AI agent feature that can carry out basic browser tasks like reservations.

ChatGPT Plus and Team clients will get access next, followed by Enterprise and then free clients. OpenAI shares that it plans to release a faster, more cost-effective version of the feature powered by a model that is smaller but just as efficient.

Also: How Gen AI means advanced customer experiences - see one bank's approach.

If you want access to the feature now but don't want to shell out the $200 per month, Google has a similar feature, also called Deep Research, that is available to all of its Gemini Advanced individuals through the Google One AI Premium plan that costs $20 per month.

Back in December, Altman even replied to an X user who asked Altman to "do a deep research feature like Gemini but improved," with "kk," suggesting that the newly released Deep Research feature is OpenAI's answer to Google.

Last week, Microsoft also presented a feature capable of more thorough reasoning called Think Deeper, which allows customers to leverage OpenAI's O1 reasoning model to deliver higher-quality responses to complex prompts. However, unlike Gemini and OpenAI's Deep Research elements, it doesn't have agentic capabilities or access to the internet. The biggest perk is that the experience is entirely free.

Life is like a box of chocolate. Generated using DALL-E.

My momma always stated "Life was like a box of chocolates. You never know what you’re gonna get...

Research FunSearch: Making new discoveries in mathematical sciences using Large Language Models Share.

Research A glimpse of the next generation of AlphaFold Share.

Progress modification: Our latest AlphaFold model demonstrates significantly impro...

The US Copyright Office's new ruling on AI art is here - and it could change everything

Last week, the US Copyright Office released its detailed analysis and comprehensive guidelines on the issue of copyright protection and AI-generated work.

For a government legal document, it is a fascinating exploration of the intersection of artificial intelligence and the very concept of authorship and creativity. The study's authors conduct a deep dive, taking in comments from the general public and experts alike, and producing an analysis of what it means to creatively author a work.

Also: How to use Microsoft Image Creator to generate and edit stunning AI images for free.

They then explore the issue of whether an AI-generated work versus an AI-assisted work is subject to copyright protection, and what that means not only for individual authors but also for the encouragement of creativity and innovation in society as a whole.

This is the second of what will be a three-part findings from the Copyright Office. Part 1, , explored digital replicas, using digital technology to "realistically replicate" someone's voice or appearance.

Part 3 is expected to be released later this year. It will focus on the issues of training AIs using copyrighted works, aspects of licensing, and how liability might be allocated in cases where a spectacular AI failure can be attributed to training (which sometimes results in litigation).

As it turns out, copyright -- or at least the protection of the rights of creators -- was considered so essential by America's founding fathers that it was listed in the Enumerated Powers clause (Article 1, Section 8) of the US Constitution.

To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the.

As a matter of priority, the powers to collect taxes and coin money were listed before the copyright clause, but declaring war, raising an army, and maintaining a navy were listed after the protection of creative rights clause.

In the minds of the framers, copyright wasn't just about the ability to collect royalties and make some cash; it was about promoting the progress of science and (and I love this) the "useful" arts. Silly arts, they didn't care about. But useful arts, those need protection. I'll leave it as an exercise for the reader to decide whether to consider blogging like I'm doing here to be "useful" or not!

Their point in protecting rights to creativity was to push progress forward, and they recognized that some creators needed incentive to do that -- basically, to be able to make a living or build a business based on their creative endeavors.

I'd love to know what Thomas Jefferson and old Ben Franklin would have made of ChatGPT!

Prior to issuing the investigation detailing the Copyright Office's determination about AI and copyrights, the agency issued a Notice of Inquiry, where they invited comments on AI-related policy issues.

A Notice of Inquiry, when properly framed and processed, is a great way for a federal agency to involve the public and gain insights from a wide range of individuals and organizations.

The agency asked five key questions that are reflected in their final determination. Those questions were:

Does the Copyright Clause in the US Constitution permit copyright protection for AI-generated material? Under copyright law, are there circumstances when a human using a generative AI system should be considered the "author" of the material produced by the system? Is legal protection for AI-generated material desirable as a policy matter? If so, should it be a form of copyright or a separate sui generis [original] right? Are any revisions to the Copyright Act necessary to clarify the human authorship requirement?

The Copyright Office received more than 10,000 comments, about half of which directly addressed the above questions. Throughout the agency's analysis, the authors refer to specific comments made by citizens in response to this Notice of Inquiry.

Also: New mystery AI image generator bests Midjourney and DALL-E 3 - how to try it.

The full research answers all the questions above, and we'll cover those answers through the rest of this article.

The Copyright Office determined, "Questions of copyrightability and AI can be resolved pursuant to existing law, without the need for legislative change."

This is a fairly essential and heavily emphasized element of the overall investigation. Basically, the question was whether new legislation would be required to incorporate the AI-related issues, or whether existing law could be applied to the new technology.

The Copyright Office maintained that the existing law has been flexible enough to incorporate new technology, having added other media and methods of creativity over the years.

Also: How to use AI to create a logo for free.

The Copyright Office also determined that, "The case has not been made for additional copyright or sui generis protection for AI-generated content." Sui generis, for those who don't have a Duolingo Latin subscription, means one of a kind or unique.

Basically, the Copyright Office doesn't believe that AI-generated copyright issues need unique legislation or protection.

Does tool use disqualify copyright protection?

The Copyright Office determined, "The use of AI tools to assist rather than stand in for human creativity does not affect the availability of copyright protection for the output."

In other words, if you choose to use a computer keyboard to write an article instead of a pen and ink, you can still copyright your writing. The issue at hand is whether the technology sufficiently separates the author from their creation such that the creation isn't human-inspired or driven.

Also: What to know about DeepSeek AI, from cost indicates to data privacy.

We'll come back to this question for more comprehensive generative AI, but the Copyright Office was clear that a tool used to help creativity (like, for example, an automatic masking tool in a video editor) is not, itself, a disqualifying factor.

Can copyright protect AI-generated material?

The Copyright Office made two related determinations here:

Copyright protects the original expression in a work created by a human author, even if the work also includes AI-generated material.

Copyright does not extend to purely AI-generated material, or material where there is insufficient human control over the expressive elements.

At the core of their determination is existing copyright law, which the Office does not believe needs to be modified for the case of generative AI.

Existing copyright law is for the benefit of humans. As a result, they believe that whatever work is copyrighted must have been substantially created by a human, not by another entity.

Also: Are ChatGPT Plus or Pro worth it? Here's how they compare to the free version.

What this means is that the bulk of the creative process must have gone through the human thought process and human activity, as opposed to the bulk of the creative process being created or generated by artificial intelligence.

Can you copyright the output of prompts?

The official title of the head of the US Copyright Office is Register of Copyrights. Way back in 1965, long before generative AI was anything more than an idea in an Isaac Asimov novel, then Register of Copyrights Abraham Kaminstein did a bit of a deep dive into considering the relationship between human authorship and machine generation. He expressed:

The crucial question appears to be whether the "work" is basically one of human authorship, with the computer merely being an assisting instrument, or whether the traditional elements of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.

Using this as precedent, the Copyright Office determined that prompts, just on their own, do not show enough human workings to be protected by copyright. The determination is, "Based on the functioning of current generally available technology, prompts do not alone provide sufficient control."

Here, the issue becomes a bit of a challenge. The Copyright Office, in the research, stated:

Where AI merely assists an author in the creative process, its use does not change the copyrightability of the output. At the other extreme, if content is entirely generated by AI, it cannot be protected by copyright. Between these boundaries, various forms and combinations of human contributions can be involved in producing AI outputs.

Copyright is something of a you'll-know-it-when-you-see-it sort of protection. This is why copyright disputes end up in court on a regular basis. There are specific factors, though. For example, in Feist Publications, Inc. v. Rural Tel. Serv. Co., 499 US 340 (1991), the US Supreme Court ruled that ideas or facts, of themselves, are not protectable by copyright law.

Additionally, the mere act of doing hard work to create something does not justify copyright. The court determined that "sweat of the brow" wasn't enough to qualify. But almost any creative effort on the part of a human does open the door to protection.

The court stated, "The requisite level of creativity is extremely low; even a slight amount will suffice. The vast majority of works make the grade quite easily, as they possess some creative spark, 'no matter how crude, humble or obvious' it might be."

Also: OpenAI launches new o3-mini model - here's how free ChatGPT consumers can try it.

In this context, the Copyright Office ruled that current law allows for determinations on whether there's a humble or crude spark of human creativity. The analysis stated, "Whether human contributions to AI-generated outputs are sufficient to constitute authorship must be analyzed on a case-by-case basis."

But what about creative choices using a given medium?

This issue was debated over a century ago, in 1884. In Burrow-Giles Lithographic business v. Sarony, 111 US 53 (1884), the US Supreme Court examined whether a machine-produced image, like a photograph, could be considered the result of human creativity.

Keep in mind that a photograph is not created by human hands. If anything, it is created by light, and electronics or chemicals. There are some optics involved as well.

The image created is the result of light during a very small fraction of a second, processed by what is essentially a machine. Unlike in 1884, our images are usually stored by computer.

The only participation of a human in a photograph is choosing where to point the camera, perhaps what lens to put on the camera, which image to present to the public, and when to take the picture. In the case of smartphone pictures, as well as many point-and-shoot standalone cameras, the human involvement is rarely more than a millimeter's flex of an index finger.

The Court ruled, however, that there were creative actions undertaken by the photographer, including posing a subject, costuming, set design, and other aspects of portraiture. For a nature photographer, control involves choosing the direction of the photograph and the time of day. For a photojournalist, it's getting to the location of the action and finding the one evocative microsecond that tells a story.

Also: This new Google AI tool lets you easily generate images from other photos - no prompt required.

The Copyright Office reflected this in its determination of authorship and creativity. The Office stated, "Human authors are entitled to copyright in their works of authorship that are perceptible in AI-generated outputs, as well as the creative selection, coordination, or arrangement of material in the outputs, or creative modifications of the outputs."

So, a prompt on its own isn't worthy of protection. Prompting mixed with creativity might be, but will be adjudicated on a case-by-case basis.

OK, fine. Let's put that to the test with a little thought experiment.

The above artwork, called Oblivious, was a project I did using Midjourney and Adobe Photoshop, which I documented in this article. Midjourney and Photoshop were, essentially, my creative medium, but the vision was my own.

I wanted to create a work evocative of Hopper's Nighthawks, but that could stand on its own. I called it Oblivious because the man quite possibly doesn't even notice that there's a giant white rabbit just a few feet to his right.

It's an allegory for the idea that we're so engrossed in our phones that we miss even the most obvious things around us. You also sense that the man would rather just tap on his phone than go in and get a nice piece of pie or a hot beverage, putting off creature comforts in favor of whatever fascinates him so much on that screen.

I also love how the rabbit represents change and newness but conveys a deep sense of longing, because he can never be inside and part of the diner milieu.

This was an image that was created based on my wanting to tell a story. It combined the AI's ability to create the graphic and my ability to guide it to what I envisioned.

Is this something where the AI did all the work and all I did was paste in some words? Or do I deserve any credit for the mood, the commentary, and the message the art shares with the observer? Why would my creativity using the medium of Midjourney count for any less than my creativity using my favorite Sony camera?

Also: How to use Gemini to generate higher-quality AI images now - for free.

One of the Copyright Office's concerns is that choosing from a variety of information or choosing from a variety of generated pictures is not creating. I would argue that a photographer does that as part of his or her craft. For example, a photographer might take 100 pictures and choose just one to submit to a magazine or for a contest.

Choosing, the act of deciding between representations, has long been part of the creative process, as I showed through the choices I documented in the article about Oblivious.

Why should choosing a photo out of hundreds or thousands of other images shot during a photoshoot be any more the act of human creativity than using Midjourney with a carefully written text prompt, getting back four variations, and choosing the best variation?

Personally, I consider Oblivious to be my work of art because it is the result of a vision that I started with and refined as I was executing the creative process.

The only difference was that instead of my medium being a brush and paint, or camera and lens, my medium was articulating to an AI what I wanted to see and how I wanted it to place things. It's still my work of art.

I think we're going to see a number of issues (and a ton of litigation) where this gray area comes into effect, where there is some question of whether a work was mostly human-authored, mostly AI-authored, or a collaboration of both human and AI.

I expect the camera analogy and all the established case law on copyrighting photographs to strongly influence future determinations of AI copyright litigation.

Also: Midjourney's AI-image generator website is now officially open to everyone - for free.

What do you think? Do you think the Copyright Office made the right determination? Do you think the result of prompting should be copyrighted? Let us know in the comments below.

Research Scaling up learning across many different robot types Share.

Together with partners from 33 academic labs, we have pooled ...

New research proposes a system to determine the relative accuracy of predictive AI in a hypothetical medical setting, and when the system should defer...

Impact MuZero, AlphaZero, and AlphaDev: Optimizing computer systems Share.

As part of our aim to build increasingly capable and gen...