This article was first published on June 8, 2023.
Over 150 years ago, statistician and founder of modern nursing, Florence Nightingale, made data visualisations that were designed to change how society behaved. She successfully took the value of data, applied it to a real-world use case of poor sanitation and overcrowding and managed to change the way we care for humans.
Although she acquired the nickname Lady with the Lamp due to her night rounds as a nurse, she was a lady with a vision; studying data before data analysis was even a term acknowledged in the medical research world. This vision was the clear articulation of data to make it digestible.
The proliferation of automation and AI applications in our everyday lives has supercharged data discussions. As the direction of AI is still emerging, it is important that we assess our data sources, reliability and future data needs.
Are we on the brink of an AI takeover? That’s the question on the minds of many technology leaders and researchers today. Everyone from Elon Musk to Steve Wozniak to doctors and public health experts is coming out in favour of AI regulation before AI continues its mission to revolutionise industries.
AI models require quality data inputs
As AI continues to revolutionise industries from healthcare to finance, the need for data to train these models has grown exponentially. But what happens when we run out of data?
AI models have risks associated with poor data inputs. Can secure blockchain data provide one solution that mitigates the effects of data shortages? This will depend on the datasets and the willingness to share data across blockchains. In order to maximise the information that the AI model can work with, collaboration amongst all stakeholders is key.
According to Hugo Philion, CEO and Co-Founder of Flare Networks, “Data that varies over time, such as stock market indices, weather and commodity prices, can be brought on-chain by the Flare Time Series Oracle in a highly decentralised way.”
“Blockchain data is by definition highly structured, following a cumulative distributed ledger structure where every new block is linked to all historical blocks via cryptographic hashes. This is what ensures the immutability of the database,” continues Philion.
Pooling multiple sources of data to create larger, more accurate datasets can make a big difference when training AI models.
“The problem isn’t necessarily how to organise existing blockchain data to make it more useful. The problem is how to bring more types of data on-chain, from other blockchains and from Web2 APIs, making them available where they are needed for Dapps to execute and provide greater utility for users,” notes Philion.
Also Read: Planning a trip: Is the future of sustainable travel in the metaverse?
Alongside the sheer size of the data sources comes the task of managing these sources and using the right inputs to extract the right information.
Using data for informed decision making
There are currently few incentives for local governments to increase efficiency. Besides self-motivation and pride of place, local governments tend to pass along responsibility for leadership strategies to the higher powers of the state. Transparency in planning, proposals, land use, regulations and infrastructure development is key to ensuring a stable local economy and establishing a community based on trust. On-chain data guarantee a store of records like never before.
Providing accurate data is a key priority for TangleHUB, a decentralised storage solution working with IOTA. However, as users can opt out of providing their data, the data inputs are difficult to predict for the development of future products and solutions to existing problems.
“With the advent of more machine learning and AI for processing data, the limitations of data within councils need to be addressed. The problem with centralised storage is that somebody has controlled access, and if the metadata isn’t encrypted, there are risks associated with the security of this data,” says Bas van Sambeek, Communications Specialist at TangleHUB.
However, using optimised data management to reduce waste and effectively allocate local budgets could provide a welcome boost to local economies and employment opportunities. If councils exercise their power to provide long-term positive outcomes for local citizens, then it could save the taxpayers millions in revenue. Also, individuals are more aware of their data usage and rights.
Unlocking the power of decentralised data
When working in the blockchain realm, there are several questions popping up in data circles, and most are concerned with the sharing of data for effective AI management. “How can we bring private data to AI, and how can we ensure that everybody involved gets their fair share of what comes out of this?” said Robin Lehmann, CEO & Co-Founder of Data Union App.
Using self-serving analytics to empower better levels of care, health, and lifestyle management is more commonplace these days as people have familiarised themselves with mobile applications that provide data on their everyday activities. Three sectors that have embraced individual data management are fitness, health and work. Examples include your Fitbit, your monthly health goals and your performance at work. In the future, this may apply to other aspects of our lives, including our relationship with public services.
“We see a huge need for people to be able to take back control over their data and for people to trust data that they see. So if you have a local council, then that data has to be absolutely reliable, or there has to be a confidence interval in that, along with that data, to be able to use it as input for the decision process,” continues Van Sambeek.
For Philion, “Personal data sovereignty will likely make engaging with services less convenient initially until the technology matures and solutions become easier to use.”
Last week, at a Crypto and AI conference, Richard Blythman, Founder of AlgoveraAI, noted the potential to add new utility with LLM frameworks. Algovera is focused on building end-to-end solutions for customised versions of LLM flows, assistants and agents. “The Crypto agent framework paired with LLM framework provides a whole lot of new utility where we can build new use cases.”
Martin Koeppelmann, Founder of Gnosis, one of the first Ethereum sidechains with over 120,000 validators, looked at data from the perspective of the AI agent. He discussed wrapping AI services on the chain for agents, “An AI agent will not have a bank account or a credit card but may very well be able to control a private key.”
By using blockchain technology as a way of tracking our data, the end user can own and control their own data. Still, also the AI will be capable of using whatever data we feed it to access services in ways that may be unimaginable today. Right now, providing the right infrastructure and data sets and setting basic standards in the way data is handled will provide pivotal guidance for humans and AI taking advantage of this data revolution.
Also Read: Celebrate World Environment Day: 4 ways blockchain and ReFi are supporting a greener future
“The lessons being learned currently are that this trust was perhaps necessary in the past in order for a service to be provided. But if new blockchain systems can remove this need, in a simple way, abstracted away by intuitive applications, then there are many advantages to having full control over one’s data,” says Philion.
Conclusion
In a recent article, the Harvard Business Review highlights a new world order that emphasizes data access. It pointed to a scenario whereby trade or data-sharing agreements between countries could become the norm in the future. The report notes that data mobility allows for “a more productive free-trade zone, where countries mutually benefit from tapping into each other’s data reservoirs.”
As AI continues to infiltrate every industry, it’s essential that technologists collaborate to find innovative solutions to tackle the challenges of data shortage. With collaboration and creativity, project leaders can ensure that AI remains a powerful tool for solving complex problems.
The marriage of blockchain technology and AI has the potential to revolutionise some of our most vital public services. Just as Florence Nightingale pioneered modern nursing through her meticulous analysis of data, we too have the opportunity to reshape the future of AI by harnessing the power of blockchain.
—
Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic
Join our e27 Telegram group, FB community, or like the e27 Facebook page
Image credit: Canva Pro
The post Running on empty: What happens when AI models run out of data? appeared first on e27.