How to ensure AI's outcomes are no longer 'subject to available data'

Without robust data collection, promises of transformative AI solutions remain empty Image: Getty Images/iStockphoto
- The potential of artificial intelligence (AI) hinges on high-quality, first-order data.
- Advanced sensors enable robots to extract massive, high-resolution datasets, creating the foundation for effective AI.
- Companies must shift to innovating technologies that generate the first-order data needed to fuel intelligent solutions.
Last year, AI dominated conversations across the globe, highlighted by OpenAI's Sam Altman's presence on the stage at the 2024 Annual Meeting in Davos, Switzerland, where he addressed the theme: "Technology in a Turbulent World."
For many, AI's sudden rise felt like an alien invasion. While it was widely acknowledged that this new technology would reshape the technological landscape – from Silicon Valley boardrooms to governmental corridors – it wasn’t clear how, when or what this new world would look like.
As the world’s biggest companies turn to technology and the 47th US President Donald Trump surrounds himself with technologists, including Elon Musk, David Sacks and Marc Andreessen, we’re about to see a whole new focus on innovation. That includes harnessing the power of AI.
Read the fine print
While there is a vast array of talent and exciting developments within the sector, this promise of AI is increasingly being undermined by a fundamental flaw: outcomes are subject to available data.
I recently spoke with a Fortune 500 CEO who described a common experience of daily pitches from AI companies, each promising to revolutionize his business, only to be met with a consistent roadblock. At the end of every presentation, a crucial asterisk appears: the requirement for data.
When it comes to the built world and the infrastructure we use every day, existing data either doesn’t exist or is limited and of poor quality. That’s because it’s usually collected by what I call “Joe on a rope” – a human, often dangling precariously, taking single readings by hand every few feet.
This manual process is prone to human error and inconsistencies, relying on them to maintain precise positioning and perform identical measurements repeatedly, often in difficult conditions.
So, when AI companies promise solutions but require existing data, you can see why it falls short. Companies don’t have the level of data that enables AI to work.
Atoms to bits
First-order data is crucial for successful AI applications, which is why Gecko Robotics, for example, entered the space to develop robotic systems that can collect high-resolution, comprehensive data from challenging environments.
Advances in the world of robotics mean we can now access and gather data on areas inaccessible to humans, such as the interiors of industrial facilities or the inner workings of a power plant.
Put simply, robots equipped with different sensors turn the physical world of atoms into bits. They can collect millions of data points on a structure, digitally map structures and then take remote actions using mobile robots.
By digitalizing these atoms, AI software can then be powered to identify defects, predict failures and optimize efficiency. Robots are the miners and the data we are collecting is digital gold.
The future belongs to companies that can harness the power of first-order data.
”A problem for the software world
First-order data isn’t just limited to fuelling AI for the built world – it applies to all industries. In fact, over the next five years, companies that don’t have access to first-order data layers to inform their software offerings will struggle against those that do.
The era of generic software solutions built on assumptions and limited data is coming to an end. The future belongs to companies that can harness the power of first-order data to build intelligent, tailored solutions that truly address the unique challenges of their customers.
To achieve this, we need to shift our focus from data consumption to data creation. Instead of relying on existing datasets that may be incomplete or biased, we need to invest in innovative technologies that can generate high-quality, first-order data.
By embracing this data-centric approach, we can unlock the full potential of AI and drive innovation across industries. We can create smarter cities, improve healthcare outcomes, optimize supply chains and develop more sustainable solutions for the future.
We need to put an end to “subject to available data” and go out there and collect it.
Don't miss any update on this topic
Create a free account and access your personalized content collection with our latest publications and analyses.
License and Republishing
World Economic Forum articles may be republished in accordance with the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International Public License, and in accordance with our Terms of Use.
The views expressed in this article are those of the author alone and not the World Economic Forum.
Stay up to date:
Artificial Intelligence
Forum Stories newsletter
Bringing you weekly curated insights and analysis on the global issues that matter.
More on Emerging TechnologiesSee all
Gideon Lichfield
February 11, 2025