Why ‘open source’ AIs could be anything but, the derailment risks of long freight trains, and breeding better wheat
Digest
The Nature Podcast delves into three intriguing scientific stories. First, Nick Petrich Howell discusses 'openwashing' in AI, where companies claim their models are open but may not meet all transparency criteria. The researchers highlight BloomZed as a truly open model, emphasizing the importance of open AI for research and regulation. Second, Dan Fox examines a study showing that longer freight trains are statistically more likely to derail, raising concerns about safety as the industry moves towards efficiency. The study uses a unique method to analyze data and highlights the need for further research. Finally, Ben explains how researchers are using ancient wheat varieties to improve modern wheat, seeking traits like nitrogen use efficiency and resistance to wheat blast. This involves extensive breeding programs and could take decades to yield results, but offers a promising avenue for enhancing a vital food source.
Outlines
Introduction
This Chapter introduces the Nature Podcast and its focus on the Nature Briefing, featuring stories from the past couple of weeks. It introduces the podcast team members, Nick Petrich Howell and Dan Fox, who will be discussing the stories.
Openwashing in AI
This Chapter explores the concept of 'openwashing' in AI, where companies claim their models are open but may not meet all transparency criteria. The researchers highlight BloomZed as a truly open model, emphasizing the importance of open AI for research and regulation. The EU's AI Act is mentioned, which may require less stringent requirements for open source AI models.
Longer Freight Trains and Derailment Risk
This Chapter examines a study showing that longer freight trains are statistically more likely to derail, raising concerns about safety as the industry moves towards efficiency. The study uses a unique method to analyze data and highlights the need for further research. The discussion touches upon the potential causes of derailment and the ongoing debate about railway safety regulations in the US.
Ancient Wheat Varieties for Modern Improvement
This Chapter explores how researchers are using ancient wheat varieties to improve modern wheat, seeking traits like nitrogen use efficiency and resistance to wheat blast. This involves extensive breeding programs and could take decades to yield results, but offers a promising avenue for enhancing a vital food source. The discussion highlights the importance of genetic diversity and the potential benefits of using historical landraces.
Keywords
Openwashing
A term used to describe the practice of companies claiming their AI models are open, while not meeting all transparency criteria. It is analogous to 'greenwashing,' where companies exaggerate their environmental credentials. Openwashing can be problematic as it can mislead researchers and regulators about the true openness of AI models.
BloomZed
An open-source AI model developed by an international and largely academic collaboration. It is highlighted as a truly open model, as its code, training data, and documentation are publicly available. BloomZed serves as an example of how AI models can be developed and shared openly, promoting transparency and collaboration in the field.
Freight Train Derailment
The risk of a freight train derailing, which is influenced by factors such as train length, weight distribution, and track conditions. A recent study has shown that longer freight trains are statistically more likely to derail, raising concerns about safety as the industry moves towards using longer trains for efficiency. This issue is particularly relevant in the US, where long freight trains are common.
Wheat Blast
A fungal disease that can severely damage wheat crops, leading to significant yield losses. Researchers are exploring the use of ancient wheat varieties to identify genes that provide resistance to wheat blast. This effort is part of a broader initiative to improve wheat's resilience to various threats, including climate change and fungal infections.
Landraces
Traditional varieties of crops that have been cultivated and adapted to specific environments over generations. Landraces often possess valuable genetic diversity, including traits that may be lost in modern varieties. Researchers are increasingly turning to landraces to identify genes that can be used to improve modern crops, such as wheat, for traits like disease resistance, drought tolerance, and nutrient efficiency.
Q&A
What is 'openwashing' in AI, and why is it a concern?
'Openwashing' refers to companies claiming their AI models are open while not meeting all transparency criteria. This is problematic because it can mislead researchers and regulators about the true openness of these models, hindering progress in AI research and regulation.
How does the length of a freight train affect the risk of derailment?
A recent study has shown that longer freight trains are statistically more likely to derail. This is due to factors such as weight distribution and the complexity of managing a train that can be over a mile long. The study highlights the need for further research and potential safety concerns as the industry moves towards using longer trains.
Why are researchers looking to ancient wheat varieties to improve modern wheat?
Researchers are exploring ancient wheat varieties, known as landraces, to identify genes that can enhance modern wheat's resilience to threats like climate change and fungal infections. These landraces often possess valuable genetic diversity that has been lost in modern varieties, offering a promising avenue for improving a vital food source.
What are some of the potential benefits of using ancient wheat varieties in modern breeding programs?
Ancient wheat varieties can provide genes for traits like nitrogen use efficiency, resistance to wheat blast, and other desirable characteristics. This could lead to more sustainable and resilient wheat crops, contributing to global food security.
What are the challenges associated with using ancient wheat varieties in modern breeding programs?
Breeding programs involving ancient wheat varieties can be time-consuming and complex. It takes time to identify and isolate desirable genes, and then to cross them into modern varieties while ensuring that no undesirable traits are introduced. This process can take decades to yield results.
Show Notes
00:31 How open are ‘open source’ AI systems?
Many of the large language models powering AI systems are described as ‘open source’ but critics say this is a misnomer, with restricted access to code and training data preventing researchers from probing how these systems work. While the definition of open source in AI models is yet to be agreed, advocates say that ‘full’ openness is crucial in efforts to make AI accountable. New research has ranked the openness of different systems, showing that despite claims of ‘openness’ many companies still don’t disclose a lot of key information.
Nature News: Not all ‘open source’ AI models are actually open: here’s a ranking
06:12 Why longer freight trains are more prone to derailment
In the US, there are no federal limits on the length of a freight train, but as companies look to run longer locomotives, questions arise about whether they are at greater risk of derailment. To find out, a team analysed data on accidents to predict the chances of longer trains coming off the tracks. They showed that replacing two 50-car freight trains with one 100-car train raises the odds of derailment by 11%, with the chances increasing the longer a train gets. While derailments are uncommon, this could change as economic pressures lead the freight industry to experiment with ever-longer trains.
Scientific American: Longer and Longer Freight Trains Drive Up the Odds of Derailment
11:44 How historic wheat could give new traits to current crops
Genes from century-old wheat varieties could be used to breed useful traits into modern crops, helping them become more disease tolerant and reducing their need for fertiliser. Researchers sequenced the genomes of hundreds of historic varieties of wheat held in a seed collection from the 1920s and 30s, revealing a huge amount of genetic diversity unseen in modern crops. Plant breeding enabled the team to identify some of the areas of the plants’ genomes responsible for traits such as nutritional content and stress tolerance. It’s hoped that in the long term this knowledge could be used to improve modern varieties of wheat.
Science: ‘Gold mine’ of century-old wheat varieties could help breeders restore long lost traits
Hosted on Acast. See acast.com/privacy for more information.