Tagged / conversation

Conversation article: Fossil footprints reveal what may be the oldest known handcarts – new research

Professor Matthew Bennett and Dr Sally Reynolds write for The Conversation about their research at White Sands National Park, where they have found the earliest evidence of prehistoric transport…

Fossil footprints reveal what may be the oldest known handcarts – new research

Matthew Robert Bennett, Bournemouth University and Sally Christine Reynolds, Bournemouth University

If you’re a parent you’ve probably tried, at some point, to navigate the supermarket with a trolley, and at least one child in tow. But our new study suggests there was an ancient equivalent, dating to 22,000 years ago. This handcart, without wheels, was used before wheeled vehicles were invented around 5,000 years ago in the Middle East.

Recently our research team discovered some remarkable fossil traces which might give a hint. These traces were found alongside some of the oldest known human footprints in the Americas at a place called White Sands in New Mexico.

In the last few years, several footprint discoveries at this site have begun to rewrite early American history – pushing back the arrival of the first people to enter this land by 8,000 years.

There is some controversy around the age (23,000 years old) of these footprints, with some researchers unhappy with our dating methods. But they provide a remarkable picture of past life on the margins of a large wetland at the end of the last ice age.

The footprints tell stories, written in mud, of how people lived, hunted and survived in this land. Footprints connect people to the past in a way that a stone tool or archaeological artefact never can. Traditional archaeology is based on the discovery of stone tools. Most people today have never made a stone tool but almost all of us will have left a footprint at some time, even if it is only on the floor of the bathroom.

Today, modern shopping trolleys can be found rusting in canals, rivers or abandoned in shrubbery. But ancient versions would have probably been of wood and simply rotted away. We know that transport technology must have existed.

Everyone has stuff to transport, but we have no record of it until written histories. At White Sands, we found drag-marks made by the ends of wooden poles while excavating for fossil footprints. Sometimes these appear as just one trace, while at other times they occur as two parallel, equidistant traces.

A pole or poles used in this fashion is called a travois. These drag-marks are preserved in dried mud that was buried by sediment and revealed by a combination of erosion and excavation. The drag-marks extend for dozens of metres before disappearing beneath overlying sediment. They clip barefoot human tracks along their length, suggesting the user dragged the travois over their own footprints as they went along.

To help interpret these features, we conducted a series of tests on mud flats both in Dorset, UK, and on the coast of Maine, US. We used different combinations of poles to recreate simple, hand-pulled travois.

In our experiments the pole-ends dragged along the mud truncate footprints in the same way as the fossil example in New Mexico. These features in the fossil examples were also always associated with lot of other human footprints travelling in a similar direction, many of which, judging by their size, were made by children.

We believe the footprints and drag-marks tell a story of the movement of resources at the edge of this former wetland. Adults pulled the simple, probably improvised travois, while a group of children tagged along to the side and behind.

The research team has benefited from the insight of the Indigenous peoples we work with at White Sands, and they interpret the marks in this way as well. We cannot discount that some of the marks may be made by dragging firewood, but this does not fit all the cases we found.

Travois are known from historical documents and accounts of Indigenous peoples and their traditions. They were more commonly associated with dogs or horses, but they were pulled by humans in our tests.

As such they represent early examples of the handcart or wheelbarrow, but without the wheel. The earliest record of a wheeled vehicle dates from Mesopotamia (modern day Iraq), in 2,500BC. We think the travois were probably improvised from tent poles, firewood and spears when the need arose.

Maybe they were created to help move camp, or more likely, transport meat from a hunting-site. In the latter context the analogy with the shopping trolley comes to the fore, as does the pained expression of the adults faces as they quest for resources with a gaggle of children in tow.The Conversation

Matthew Robert Bennett, Professor of Environmental and Geographical Sciences, Bournemouth University and Sally Christine Reynolds, Associate Professor in Hominin Palaeoecology, Bournemouth University

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Conversation article: How a New York Times copyright lawsuit against OpenAI could potentially transform how AI and copyright work

Professor Dinusha Mendis writes for The Conversation about the potential copyright implications of AI as a lawsuit is lodged by the New York Times against the creator of ChatGPT…

How a New York Times copyright lawsuit against OpenAI could potentially transform how AI and copyright work

Stas Malyarevsky / Shutterstock

Dinusha Mendis, Bournemouth University

On December 27, 2023, the New York Times (NYT) filed a lawsuit in the Federal
District Court in Manhattan against Microsoft and OpenAI, the creator of ChatGPT,
alleging that OpenAI had unlawfully used its articles to create artificial intelligence (AI) products.

Citing copyright infringement and the importance of independent journalism to democracy, the newspaper further alleged that even though the defendant, OpenAI, may have “engaged in wide scale copying from many sources, they gave Times content particular emphasis” in training generative artificial intelligence (GenAI) tools such as Generative Pre-Trained Transformers (GPT). This is the kind of technology that underlies products such as the AI chatbot ChatGPT.

The complaint by the New York Times states that OpenAI took millions of copyrighted news articles, in-depth investigations, opinion pieces, reviews, how-to guides and more in an attempt to “free ride on the Times’s massive investment in its journalism”.

In a blog post published by OpenAI on January 8, 2024, the tech company responded to the allegations by emphasising its support of journalism and partnerships with news organisations. It went on to say that the “NYT lawsuit is without merit”.

In the months prior to the complaint being lodged by the New York Times, OpenAI had entered into agreements with large media companies such as Axel-Springer and the Associated Press, although notably, the Times failed to reach an agreement with the tech company.

The NYT case is important because it is different to other cases involving AI and copyright, such as the case brought by the online photo library Getty Images against the tech company Stability AI earlier in 2023. In this case, Getty Images alleged that Stability AI processed millions of copyrighted images using a tool called Stable Diffusion, which generates images from text prompts using AI.

The main difference between this case and the New York Times one is that the newspaper’s complaint highlighted actual outputs used by OpenAI to train its AI tools. The Times provided examples of articles that were reproduced almost verbatim.

Use of material

The defence available to OpenAI is “fair use” under the US Copyright Act 1976, section 107. This is because the unlicensed use of copyright material to train generative AI models can serve as a “transformative use” which changes the original material. However, the complaint from the New York Times also says that their chatbots bypassed the newspaper’s paywalls to create summaries of articles.

Even though summaries do not infringe copyright, their use could be used by the New York Times to try to demonstrate a negative commercial impact on the newspaper – challenging the fair use defence.

ChatGPT
Giulio Benzin / Shutterstock

This case could ultimately be settled out of court. It is also possible that the Times’ lawsuit was more a negotiating tactic than a real attempt to go all the way to trial. Whichever way the case proceeds, it could have important implications for both traditional media and AI development.

It also raises the question of the suitability of current copyright laws to deal with AI. In a submission to the House of Lords communications and digital select committee on December 5, 2023, OpenAI claimed that “it would be impossible to train today’s leading AI models without copyrighted materials”.

It went on to say that “limiting training data to public domain books and drawings created more than a century ago might yield an interesting experiment but would not provide AI systems that meet the needs of today’s citizens”.

Looking for answers

The EU’s AI Act –- the world’s first AI Act –- might give us insights into some future directions. Among its many articles, there are two provisions particularly relevant to copyright.

The first provision titled, “Obligations for providers of general-purpose AI
models” includes two distinct requirements related to copyright. Section 1(C)
requires providers of general-purpose AI models to put in place a policy to respect EU copyright law.

Section 1(d) requires providers of general purpose AI systems to draw up and make publicly available a detailed summary about content used for training AI systems.

While section 1(d) raises some questions, section 1(c) makes it clear that any use of copyright protected content requires the authorisation of the rights holder concerned unless relevant copyright exceptions apply. Where the rights to opt out has been expressly reserved in an appropriate manner, providers of general purpose AI models, such as OpenAI, will need to obtain authorisation from rights holders if they want to carry out text and data mining on their copyrighted works.

Even though the EU AI Act may not be directly relevant to the New York Times complaint against OpenAI, it illustrates the way in which copyright laws will be designed to deal with this fast-moving technology. In future, we are likely to see more media organisations adopting this law to protect journalism and creativity. In fact, even before the EU AI Act was passed, the New York Times blocked OpenAI from trawling its content. The Guardian followed suit in September 2023 – as did many others.

However, the move did not allow material to be removed from existing training
data sets. Therefore, any copyrighted material used by the training models up until then would have been used in OpenAI’s outputs –- which led to negotiations between the New York Times and OpenAI breaking down.

With laws such as those in the EU AI Act now placing legal obligations on general purpose AI models, their future could look more constrained in the way that they use copyrighted works to train and improve their systems. We can expect other jurisdictions to update their copyright laws reflecting similar provisions to that of the EU AI Act in an attempt to protect creativity. As for traditional media, ever since the rise of the internet and social media, news outlets have been challenged in drawing readers to their sites and generative AI has simply exacerbated this issue.

This case will not spell the end of generative AI or copyright. However, it certainly raises questions for the future of AI innovation and the protection of creative content. AI will certainly continue to grow and develop and we will continue to see and experience its many benefits. However, the time has come for policymakers to take serious note of these AI developments and update copyright laws, protecting creators in the process.The Conversation

Dinusha Mendis, Professor of Intellectual Property and Innovation Law; Director Centre for Intellectual Property Policy and Managament (CIPPM), Bournemouth University, Bournemouth University

This article is republished from The Conversation under a Creative Commons license. Read the original article.

CQR Lunchtime Conversations Kick Off 3rd Oct at 1 pm RLH 201

The engaging CQR lunchtime Go Create!

seminar series for 2018-19 begins with

Liz Norton, Caroline Ellis-Hill &

Ann Hemingway

“Creative ways of data gathering &

dissemination”

Oct 3rd 1-2 pm RLH 201

Come prepared for informal conversation, sharing, and audience participation!

“We will be VERY informal!”

See you there!

R