April marks the beginning of a new season for the Innovators Club with Albert Cañigueral as the guest speaker. The speaker defines himself as an explorer, writer, and consultant focused on the social and economic impacts of digital innovations shaping our era. His work revolves around research, inspiration, and guidance for companies, organizations, and individuals to adapt to the digital age and the responsible and ethical use of digital technologies. Currently, he is involved in technology transfer and ecosystem development for the Aina project at the Barcelona Supercomputing Center and collaborates with the .CAT Foundation in transitioning towards a fair digital society.
Under the title “Will AI take our jobs?”, Albert reflects on Artificial Intelligence (AI) and the future of work. There is much noise surrounding AI, often filled with myths that do not align with reality. Our relationship with machines is often portrayed as confrontational when it should be viewed as collaborative. AI is a valuable support tool within our reach; it is not meant to replace our work entirely. Workers can integrate AI into their daily routines as a good co-pilot, but with caution not to turn it into an autopilot.
What have been the key milestones in the development and evolution of Artificial Intelligence? Where are we now?
Various forms of AI have been under development for a long time, operating for decades in many areas of our daily lives. From the filters we see on YouTube or social networks, to the ads that appear on websites, machine translations, travel time and traffic estimates on Google Maps, or the banking analysis to grant us credit or not.
What has happened now is the popularization of “generative” AI and, moreover, it has been made available to millions of people through tools like GitHub Copilot (code generation), Mid Journey (graphic images), or ChatGPT and Perplexity. There has been a great deal of investment and an explosion of tools of all kinds.
A lot of hype has been generated around all of this. The consulting firm Gartner placed generative AI at the top of the “hype cycle” in late 2023.
What is a significant challenge when considering the integration of Artificial Intelligence in workplace environments?
The main challenge is understanding where AI can truly add value in the productive process of creating a product or service. We must overcome the idea of replacing people solely to reduce costs and, moreover, assess whether automation (which may be imperfect due to the nature of AI) actually helps workers.
I recommend the ILO report (International Labour Organization) on the future of work and generative AI. The challenge is to pause and ask all the important questions in the face of the urgency and FOMO (Fear Of Missing Out) associated with the immediate incorporation of AI. AI in workplace environments must be effective, deserving of our trust, and provide security for everyone (companies, workers, and society at large).
Do you believe the idea that while AI provides statistical information, humans contribute logical context is valid?
Yes, the idea is valid. One of the most well-known critics of AI, Gary Marcus, talks about having “deep learning but not deep understanding.”
With current systems, as explained in Prediction Machines (a highly recommended book), reducing the cost of making predictions with a high degree of accuracy is already a significant advancement. However, we must place these results in logical context (human interpretation) and consider how they should be interpreted when making decisions (and automating them in some cases).
If we want AI to understand the world, we need to move towards other technical architectures (more piecewise combination and less “black box magic”). There are few people researching in directions other than making current models larger. “A bigger tree will not take us to the moon” (Gary Marcus in the book Rebooting AI).
Does AI lack the ability to understand human knowledge and reality?
Up until now, there hasn’t been much progress in this capacity for understanding the world. AI lacks the concept of cause and effect. It can detect correlations on a large scale but not causality.
It also cannot understand the production process behind a result (whether it’s an image from Midjourney or an automatic translation). It can replicate results but not the process.
AI doesn’t know how to respond to things it hasn’t seen during training, and the dangerous part is that it often tries to provide an answer that may sound plausible but lacks any logical validation.
Imagine a future scenario in industrial laundry + AI. What could it be?
This is the type of question we should be asking ourselves. Drawing scenarios (with various roles and weights between the roles of people and machines), conducting cost analyses, evaluating the quality of the work of the people involved in the operation, etc.
• AI applied to the energy and water consumption of the machines?
• AI and robotics applied to process automation?
• AI to redesign the machines themselves and the entire operations process of an industrial laundry?
• etc…
I don’t dare to answer, but it seems like a good opportunity for a fresh start (from scratch) for industrial laundry sector.
share