AI firms say they can’t respect copyright. These researchers tried.; The Washington Post, June 5, 2025

Analysis by Nitasha Tiku 

with research by 

Andrea Jiménez

 and 

Will Oremus

, The Washington Post; AI firms say they can’t respect copyright. These researchers tried.

"A group of more than two dozen AI researchers have found that they could build a massive eight-terabyte dataset using only text that was openly licensed or in public domain. They tested the dataset quality by using it to train a 7 billion parameter language model, which performed about as well as comparable industry efforts, such as Llama 2-7B, which Meta released in 2023.

A paper published Thursday detailing their effort also reveals that the process was painstaking, arduous and impossible to fully automate.

The group built an AI model that is significantly smaller than the latest offered by OpenAI’s ChatGPT or Google’s Gemini, but their findings appear to represent the biggest, most transparent and rigorous effort yet to demonstrate a different way of building popular AI tools.

That could have implications for the policy debate swirling around AI and copyright.

The paper itself does not take a position on whether scraping text to train AI is fair use.

That debate has reignited in recent weeks with a high-profile lawsuit and dramatic turns around copyright law and enforcement in both the U.S. and U.K."

 

‘AI Jesus’ avatar tests man’s faith in machines and the divine; AP, November 28, 2024

 JAMEY KEATEN, AP; ‘AI Jesus’ avatar tests man’s faith in machines and the divine

"Researchers and religious leaders on Wednesday released findings from a two-month experiment through art in a Catholic chapel in Switzerland, where an avatar of “Jesus” on a computer screen — tucked into a confessional — took questions by visitors on faith, morality and modern-day woes, and offered responses based on Scripture...

Philipp Haslbauer, an IT specialist at the Lucerne University of Applied Sciences and Arts who pulled together the technical side of the project, said the AI responsible for taking the role of “AI Jesus” and generating responses was GPT-4o by OpenAI, and an open-source version of the company’s Whisper was used for speech comprehension. 

An AI video generator from Heygen was used to produce voice and video from a real person, he said. Haslbauer said no specific safeguards were used “because we observed GPT-4o to respond fairly well to controversial topics.”

Visitors broached many topics, including true love, the afterlife, feelings of solitude, war and suffering in the world, the existence of God, plus issues like sexual abuse cases in the Catholic Church or its position on homosexuality.

Most visitors described themselves as Christians, though agnostics, atheists, Muslims, Buddhists and Taoists took part too, according to a recap of the project released by the Catholic parish of Lucerne.

About one-third were German speakers, but “AI Jesus” — which is conversant in about 100 languages — also had conversations in languages like Chinese, English, French, Hungarian, Italian, Russian and Spanish."

Rethinking natural hazards research and engagement to include co-creation with Indigenous communities; npj Natural Hazards, November 22, 2024

Thomas J. Jones, Harry Nyce Jr., 

Yannick Le Moigne, 

Glyn Williams-Jones & 

Deanna Nyce 

, npj Natural Hazards ; Rethinking natural hazards research and engagement to include co-creation with Indigenous communitiesvolume

"Indigenous peoples are widely affected by natural hazards and their history and knowledge can directly inform on past events and mitigation strategies. Here we show how effective co-creation of resources and bi-lateral knowledge exchange between natural hazard researchers and local Indigenous communities provides an effective, equitable, and sustainable way to conduct research.

Introduction

Internationally there is no single agreed definition of Indigenous peoples, and here we use Indigenous as an all-encompassing international term (in Canada this includes First Nations, Métis and Inuit). The United Nations Office for Disaster Risk Reduction (UNDRR) reports1 that 476 million people in more than 90 countries identify as Indigenous and ~20% of the Earth is covered by Indigenous territories. Consequently, Indigenous peoples globally live at risk from natural hazards (e.g., volcanic eruptions, landslides, earthquakes) and also receive benefits from living in active geological areas (e.g., fertile soils, tourism, geothermal power). According to 2021 Canadian census data2, in the Province of British Columbia (B.C.) there are 290,210 people who identify as Indigenous, with 180,085, 97,865 and 1725 people self-identifying as First Nations, Métis and Inuit, respectively. All these people are susceptible to natural hazards and, as illustrated by Fig. 1, the largest volcanic eruptions, earthquakes, landslides, wildfires, and floods in British Columbia, Canada have all affected Indigenous territories. Thus, given the global spatial overlap between Indigenous peoples and natural hazards, and the need for meaningful collaboration, both research and Indigenous local knowledge should be shared for mutual benefit. Here, we outline how the Nisg̱a’a First Nation and volcanology researchers have initiated and maintained a fruitful collaboration with bi-lateral knowledge exchange and resource co-creation. Furthermore, this comment article is co-written by non-Indigenous volcanology researchers based at universities (Jones and Williams-Jones) and a government organisation (Le Moigne) and by Indigenous scholars of the Nisg̱a’a First Nation, based at a post-secondary education establishment (Nyce and Nyce Jr.)...

There are many different resources that can be co-created between researchers and Indigenous communities. For the purposes of disseminating knowledge and history about natural hazards and past events, resource examples include information boards, posters, school activities, museum displays, online media, artwork, and audio tours. Irrespective of the resource type, co-creation and development has multiple benefits. The knowledge exchange is bi-lateral, and in many cases, Indigenous Knowledge directly informs on risk mitigation and hazard perception4...

Reciprocal benefits

Based on our experience, we highlight several mutually beneficial exchanges involving both Indigenous communities and research scientists working on natural hazards. These reciprocal benefits are applicable globally and are not specific to the natural hazard or the individual researcher-Indigenous community relationship."

AI is complicating plagiarism. How should scientists respond?; Nature, July 30, 2024

 Diana Kwon , Nature; AI is complicating plagiarism. How should scientists respond?

"From accusations that led Harvard University’s president to resign in January, to revelations in February of plagiarized text in peer-review reports, the academic world has been roiled by cases of plagiarism this year.

But a bigger problem looms in scholarly writing. The rapid uptake of generative artificial intelligence (AI) tools — which create text in response to prompts — has raised questions about whether this constitutes plagiarism and under what circumstances it should be allowed. “There’s a whole spectrum of AI use, from completely human-written to completely AI-written — and in the middle, there’s this vast wasteland of confusion,” says Jonathan Bailey, a copyright and plagiarism consultant based in New Orleans, Louisiana.

Generative AI tools such as ChatGPT, which are based on algorithms known as large language models (LLMs), can save time, improve clarity and reduce language barriers. Many researchers now argue that they are permissible in some circumstances and that their use should be fully disclosed.

But such tools complicate an already fraught debate around the improper use of others’ work. LLMs are trained to generate text by digesting vast amounts of previously published writing. As a result, their use could result in something akin to plagiarism — if a researcher passes off the work of a machine as their own, for instance, or if a machine generates text that is very close to a person’s work without attributing the source. The tools can also be used to disguise deliberately plagiarized text, and any use of them is hard to spot. “Defining what we actually mean by academic dishonesty or plagiarism, and where the boundaries are, is going to be very, very difficult,” says Pete Cotton, an ecologist at the University of Plymouth, UK."

Machine ‘Unlearning’ Helps Generative AI ‘Forget’ Copyright-Protected and Violent Content; UT News, The University of Texas at Austin, May 21, 2024

UT News, The University of Texas at Austin ; Machine ‘Unlearning’ Helps Generative AI ‘Forget’ Copyright-Protected and Violent Content

"When people learn things they should not know, getting them to forget that information can be tough. This is also true of rapidly growing artificial intelligence programs that are trained to think as we do, and it has become a problem as they run into challenges based on the use of copyright-protected material and privacy issues.

To respond to this challenge, researchers at The University of Texas at Austin have developed what they believe is the first “machine unlearning” method applied to image-based generative AI. This method offers the ability to look under the hood and actively block and remove any violent images or copyrighted works without losing the rest of the information in the model.

“When you train these models on such massive data sets, you’re bound to include some data that is undesirable,” said Radu Marculescu, a professor in the Cockrell School of Engineering’s Chandra Family Department of Electrical and Computer Engineering and one of the leaders on the project. “Previously, the only way to remove problematic content was to scrap everything, start anew, manually take out all that data and retrain the model. Our approach offers the opportunity to do this without having to retrain the model from scratch.”"

Canada moves to protect coral reef that scientists say ‘shouldn’t exist’; The Guardian, March 15, 2024

Leyland Cecco, The Guardian; Canada moves to protect coral reef that scientists say ‘shouldn’t exist’

"For generations, members of the Kitasoo Xai’xais and Heiltsuk First Nations, two communities off the Central Coast region of British Columbia, had noticed large groups of rockfish congregating in a fjord system.

In 2021, researchers and the First Nations, in collaboration with the Canadian government, deployed a remote-controlled submersible to probe the depths of the Finlayson Channel, about 300 miles north-west of Vancouver.

On the last of nearly 20 dives, the team made a startling discovery – one that has only recently been made public...

The discovery marks the latest in a string of instances in which Indigenous knowledge has directed researchers to areas of scientific or historic importance. More than a decade ago, Inuk oral historian Louie Kamookak compared Inuit stories with explorers’ logbooks and journals to help locate Sir John Franklin’s lost ships, HMS Erebus and HMS Terror. In 2014, divers located the wreck of the Erebus in a spot Kamookak suggested they search, and using his directions found the Terror two years later."

Researchers tested leading AI models for copyright infringement using popular books, and GPT-4 performed worst; CNBC, March 6, 2024

Hayden Field, CNBC; Researchers tested leading AI models for copyright infringement using popular books, and GPT-4 performed worst

"The company, founded by ex-Meta researchers, specializes in evaluation and testing for large language models — the technology behind generative AI products.

Alongside the release of its new tool, CopyrightCatcher, Patronus AI released results of an adversarial test meant to showcase how often four leading AI models respond to user queries using copyrighted text.

The four models it tested were OpenAI’s GPT-4, Anthropic’s Claude 2, Meta’s Llama 2 and Mistral AI’s Mixtral.

“We pretty much found copyrighted content across the board, across all models that we evaluated, whether it’s open source or closed source,” Rebecca Qian, Patronus AI’s cofounder and CTO, who previously worked on responsible AI research at Meta, told CNBC in an interview.

Qian added, “Perhaps what was surprising is that we found that OpenAI’s GPT-4, which is arguably the most powerful model that’s being used by a lot of companies and also individual developers, produced copyrighted content on 44% of prompts that we constructed.”"

Study: Digital watermarks and AI will expedite copyright cases; Scottish Legal News, February 5, 2024

Scottish Legal News; Study: Digital watermarks and AI will expedite copyright cases

"The study was carried out by Professor James Griffin from the University of Exeter Law School and others. Researchers applied an existing AI system to copyright case law, to see how it could read and understand cases and produce outcomes in disputes concerning 3D printing. They found more complex watermarks will lead to faster and more accurate resolutions."

Editors quit top neuroscience journal to protest against open-access charges; Nature, April 21, 2023

Katharine Sanderson, Nature; Editors quit top neuroscience journal to protest against open-access charges

"More than 40 editors have resigned from two leading neuroscience journals in protest against what the editors say are excessively high article-processing charges (APCs) set by the publisher. They say that the fees, which publishers use to cover publishing services and in some cases make money, are unethical. The publisher, Dutch company Elsevier, says that its fees provide researchers with publishing services that are above average quality for below average price. The editors plan to start a new journal hosted by the non-profit publisher MIT Press.

The decision to resign came about after many discussions among the editors, says Stephen Smith, a neuroscientist at the University of Oxford, UK, and editor-in-chief of one of the journals, NeuroImage. “Everyone agreed that the APC was unethical and unsustainable,” says Smith, who will lead the editorial team of the new journal, Imaging Neuroscience, when it launches.

The 42 academics who made up the editorial teams at NeuroImage and its companion journal NeuroImage: Reports announced their resignations on 17 April. The journals are open access and require authors to pay a fee for publishing services. The APC for NeuroImage is US$3,450; NeuroImage: Reports charges $900, which will double to $1,800 from 31 May. Elsevier, based in Amsterdam, says that the APCs cover the costs associated with publishing an article in an open-access journal, including editorial and peer-review services, copyediting, typesetting archiving, indexing, marketing and administrative costs. Andrew Davis, Elsevier’s vice-president of corporate communications, says that NeuroImage’s fee is less than that of the nearest comparable journal in its field, and that the publisher’s APCs are “set in line with our policy [of] providing above average quality for below average price”."

The government dropped its case against Gang Chen. Scientists still see damage done; WBUR, February 16, 2022

Max Larkin, WBUR ; The government dropped its case against Gang Chen. Scientists still see damage done

"When federal prosecutors dropped all charges against MIT professor Gang Chen in late January, many researchers rejoiced in Greater Boston and beyond.

Chen had spent the previous year fighting charges that he had lied and omitted information on U.S. federal grant applications. His vindication was a setback for the "China Initiative," a controversial Trump-era legal campaign aimed at cracking down on the theft of American research and intellectual property by the Chinese government.

Researchers working in the United States say the China Initiative has harmed both their fellow scientists and science itself — as a global cooperative endeavor. But as U.S.-China tensions remain high, the initiative remains in place."

Massive open index of scholarly papers launches; Nature, January 24, 2022

Dalmeet Singh Chawla , Nature; Massive open index of scholarly papers launches

"An ambitious free index of more than 200 million scientific documents that catalogues publication sources, author information and research topics, has been launched.

The index, called OpenAlex after the ancient Library of Alexandria in Egypt, also aims to chart connections between these data points to create a comprehensive, interlinked database of the global research system, say its founders. The database, which launched on 3 January, is a replacement for Microsoft Academic Graph (MAG), a free alternative to subscription-based platforms such as Scopus, Dimensions and Web of Science that was discontinued at the end of 2021."

A Little-Known Statute Compels Medical Research Transparency. Compliance Is Pretty Shabby.; On The Media, April 21, 2021

 On The Media; A Little-Known Statute Compels Medical Research Transparency. Compliance Is Pretty Shabby.

"Evidence-based medicine requires just that: evidence. Access to the collective pool of knowledge produced by clinical trials is what allows researchers to safely and effectively design future studies. It's what allows doctors to make the most informed decisions for their patients.

Since 2007, researchers have been required by law to publish the findings of any clinical trial with human subjects within a year of the trial's conclusion. Over a decade later, even the country's most well-renown research institutions sport poor reporting records. This week, Bob spoke with Charles Piller, an investigative journalist at Science Magazine who's been documenting this dismal state of affairs since 2015. He recently published an op-ed in the New York Times urging President Biden to make good on his 2016 "promise" to start withholding funds to force compliance."

A New Copyright Office Warehouse–25 Years in the Making; Library of Congress, August 19, 2020

Holland Gormley, Library of Congress; A New Copyright Office Warehouse–25 Years in the Making

"The following is a guest post by Paul Capel, Supervisory Records Management Section Head.

The United States Copyright Office holds the most comprehensive collection of copyright records in the world. The Office has over 200,000 boxes of deposit copies spread among three storage facilities in Landover, Maryland; a contracted space in Pennsylvania; and the National Archives and Records Administration (NARA) facility in Massachusetts. Even with these three warehouses, that’s not enough space. Each day, the Office receives new deposits, and despite the increase in electronic deposits, our physical deposits continue to grow year after year.

These deposits are managed by the Deposit Copies Storage Unit, a dedicated team that springs into action to retrieve deposits when requested by examiners or researchers or for litigation cases. In this type of work, speed and efficiency of retrieval are critical. Managing deposits across three storage locations can present a challenge to our ideal retrieval times. When our records are stored in several locations, the potential for miscommunication or misplaced deposits increases.

This October, the Office will be opening a new 40,000 square foot warehouse that has been in discussion for over twenty-five years. We will be moving our deposits out of facilities that are more than forty years old to centrally locate them in a new state-of-the-art facility. This is a huge undertaking, and we are aiming to move 88,000 boxes from Landover in under 45 days. The new space is environmentally controlled and meets preservation requirements for the storage of federal records. Even more importantly, the new facility will allow the Office to maintain control over all our records in a single location, which will improve our retrieval times and will enable us to serve our stakeholders better.

This new facility is a great start, but we have an even bigger vision for our deposits. To truly inventory and track our deposits, the Office is investigating a warehouse management system that will help staff inventory, track, locate, and manage all the items in our warehouse. This type of system will help the Office enhance the availability and accessibility of materials, decreasing manual processing, and allowing for real-time tracking of deposits at any given time. It will also let us know who has them and when their period of retention ends.

This system will provide all the notifications  expected from any modern delivery service. Copyright Office staff will be able to obtain a copy of their order and tell when it is in transit, know when it has been delivered, and sign for it digitally. This system will also provide transparency to others who might have an interest in requesting the same deposit, to see where it currently is, who has it, and how long they have had it."

Open-access Plan S to allow publishing in any journal; Nature, July 16, 2020

Richard Van Noorden, Nature; Open-access Plan S to allow publishing in any journal

Funders will override policies of subscription journals that don’t let scientists share accepted manuscripts under open licence.

"Funding agencies behind the radical open-access (OA) initiative Plan S have announced a policy that could make it possible for researchers to bypass journals’ restrictions on open publishing. The change could allow scientists affected by Plan S to publish in any journal they want — even in subscription titles, such as Science, that haven’t yet agreed to comply with the scheme.

Plan S, which kicks in from 2021, aims to make scientific and scholarly works free to read and reproduce as soon as they are published. Research funders that have signed up to it include the World Health Organization, Wellcome in London, the Bill & Melinda Gates Foundation in Seattle, Washington, and 17 national funders, mostly in Europe. The European Commission also says it will follow the plan.

Under the initiative, scientists funded by Plan S agencies must publish their work OA. If a journal doesn’t allow that, researchers can instead post an accepted version of their article — an author accepted manuscript, or AAM — in an online repository as soon as their paper appears. This kind of author-initiated sharing is sometimes called green open access. Under Plan S, it comes with a key condition that has so far been anathema to many subscription journals: the AAM must be shared under a liberal ‘CC-BY’ publishing licence that would allow others to republish and translate the work."

Over 24,000 coronavirus research papers are now available in one place; MIT Technology Review, March 16, 2020

Karen Hao, MIT Technology Review; Over 24,000 coronavirus research papers are now available in one place

The data set aims to accelerate scientific research that could fight the Covid-19 pandemic.

"The news: Today researchers collaborating across several organizations released the Covid-19 Open Research Dataset (CORD-19), which includes over 24,000 research papers from peer-reviewed journals as well as sources like bioRxiv and medRxiv (websites where scientists can post non-peer-reviewed preprint papers). The research covers SARS-CoV-2 (the scientific name for the coronavirus), Covid-19 (the scientific name for the disease), and the coronavirus group. It represents the most extensive collection of scientific literature related to the ongoing pandemic and will continue to update in real time as more research is released."

Rumored executive order would change landscape of UC subscription partnerships; The Daily Californian, January 30, 2020

Alexandra Casey, The Daily Californian; Rumored executive order would change landscape of UC subscription partnerships

"Prominent Nobel laureate and chief scientific officer of New England Biolabs Rich Roberts has no online access to a paper he co-authored because his institution lacks a subscription to academic journal Nature Microbiology.

Roberts is one of 21 American Nobel laureates who submitted an open letter to President Donald Trump on Monday urging him to approve a rumored plan to make federally funded research free of cost and immediately accessible after publication. UC Berkeley’s Randy Schekman, who founded eLife — an open access scientific journal — led the Nobel laureates in their letter...

“This would effectively nationalize the valuable American intellectual property that we produce and force us to give it away to the rest of the world for free,” according to the letter from the publishers. “This risks reducing exports and negating many of the intellectual property protections the Administration has negotiated with our trading partners.”

The letter added that the cost shift could place an “additional burden” on taxpayers and undermine both the marketplace and American innovation."

Libraries will champion an open future for scholarship; Pittsburgh Post-Gazette, January 29, 2020

Keith Webster, Pittsburgh Post-Gazette;

Libraries will champion an open future for scholarship

Open access deals help make knowledge and education accessible to the working class

"All of us who work in academic libraries here in Pittsburgh and around the world aspire to improve the quality of science and scholarship. It’s increasingly clear that this can best be done through the open exchange of ideas and data, which can accelerate the pace and reach of scientific discovery.

The desire of researchers and their funders to make their research freely available to all is evident. As a result, the acceptance of open access publishing and article sharing services has soared in recent years. Meanwhile, the rapidly escalating journal costs experienced by libraries over the past 25 years are agreed to be unsustainable. It is against this backdrop that Carnegie Mellon University is establishing open access agreements with top journal publishers, with a special focus on the the fields of science and computing."

A face-scanning algorithm increasingly decides whether you deserve the job; The Washington Post, October 22, 2019

Drew Harwell, The Washington Post; A face-scanning algorithm increasingly decides whether you deserve the job 

HireVue claims it uses artificial intelligence to decide who’s best for a job. Outside experts call it ‘profoundly disturbing.’

"“It’s a profoundly disturbing development that we have proprietary technology that claims to differentiate between a productive worker and a worker who isn’t fit, based on their facial movements, their tone of voice, their mannerisms,” said Meredith Whittaker, a co-founder of the AI Now Institute, a research center in New York...

Loren Larsen, HireVue’s chief technology officer, said that such criticism is uninformed and that “most AI researchers have a limited understanding” of the psychology behind how workers think and behave...

“People are rejected all the time based on how they look, their shoes, how they tucked in their shirts and how ‘hot’ they are,” he told The Washington Post. “Algorithms eliminate most of that in a way that hasn’t been possible before.”...

HireVue’s growth, however, is running into some regulatory snags. In August, Illinois Gov. J.B. Pritzker (D) signed a first-in-the-nation law that will force employers to tell job applicants how their AI-hiring system works and get their consent before running them through the test. The measure, which HireVue said it supports, will take effect Jan. 1."

Wary of Chinese Espionage, Houston Cancer Center Chose to Fire 3 Scientists; The New York Times, April 22, 2019

Mihir Zaveri, The New York Times; Wary of Chinese Espionage, Houston Cancer Center Chose to Fire 3 Scientists

"“A small but significant number of individuals are working with government sponsorship to exfiltrate intellectual property that has been created with the support of U.S. taxpayers, private donors and industry collaborators,” Dr. Peter Pisters, the center’s president, said in a statement on Sunday.

“At risk is America’s internationally acclaimed system of funding biomedical research, which is based on the principles of trust, integrity and merit.”

The N.I.H. had also flagged two other researchers at MD Anderson. One investigation is proceeding, the center said, and the evidence did not warrant firing the other researcher.

The news of the firings was first reported by The Houston Chronicle and Science magazine.

The investigations began after Francis S. Collins, the director of the National Institutes of Health, sent a letter in August to more than 10,000 institutions the agency funds, warning of “threats to the integrity of U.S. biomedical research.”"

Do You Have Concerns about Plan S? Then You Must be an Irresponsible, Privileged, Conspiratorial Hypocrite; The Scholarly Kitchen, November 26, 2018

Rick Anderson, The Scholarly Kitchen; Do You Have Concerns about Plan S? Then You Must be an Irresponsible, Privileged, Conspiratorial Hypocrite

"Ultimately, though, what is most concerning about Plan S is not the behavior of those hell-bent on defending it by any means necessary. That’s just par for the course. More important is the way in which researchers themselves — the people whose work and whose freedom to choose will be directly affected by its implementation — seem to have been excluded from the process of formulating it. This shouldn’t be surprising, I guess, given the disdain in which authors and researchers are apparently held by Plan S’s creators. After all, as Science Europe’s Robert-Jan Smits puts it: “Why do we need Plan S? Because researchers are irresponsible.”

There you have it. The freedom to choose how to publish isn’t for everyone; it’s only for those who are “responsible” — which is to say, those who agree with Plan S."