S3P T&BD
Jan Jacobs & Annemiek Canjels
Fontys Green Tech Lab & Province Limburg, The Netherlands
Dr Jan Jacobs, former professor at the Fontys University of Applied Sciences and senior researcher at the Fontys Green Tech Lab, is making a point: We may have become overwhelmed by the seemingly endless possibilities of Artificial Intelligence – only to forget that nature has taught us that some of its best solutions are not a result of reasoning, but of the so-called sense-actuate response in which the real world is directly participating. In simple words: We can decide to turn up the heater because we feel cold, or we can ask a computer to analyse a million responses to various situations by peers and thus decide on our needs for us. But the result of the latter, rather indirect approach, is probably not as good and may have some other disadvantages also.
Personally, I could not agree more. In a sustainable world, we should avoid big data storage and use of the Internet of Things for nonsensical activity that only numbs our human intelligence. Yes I can still switch on a lightbulb myself. Without thinking. But I am not an AI, expert, I am just a senior policy maker that was involved as partner coordinator in INTERREG EUROPE project Regions4Food.
Jan is an experienced professional on AI/Deep Learning/Machine Learning projects. In the article below, Jan shows us examples of how most agriculture sustainability and productivity can be accomplished best by engineering the sense-actuate response in a direct machine – agri-operator relationship. It is this kind of innovation challenges that seems to get little support from the European Commission in the years to come, in spite of the urgent need for high impact solutions. Jan and I urge all agriculture policy makers to give way to the engineers.
Just few weeks ago, the European Declaration on Digital Rights and Principles was signed by the European Commission, the European Parliament and the European Council. With this Declaration, the Commission wants to ensure that people are empowered to fully enjoy the opportunities that the digital decade brings. The Declaration proposes a set of European digital rights and principles that reflect EU values and promote a sustainable, human-centric vision for the digital transformation. Europe strongly advocates a Digital Transition, which makes sense, when we look at the figures estimated: the value of the data economy in the EU-27 is expected to grow from €301 billion to €829 billion, by 2025, 60 million new jobs worldwide could be created by AI and robotics and the number of EU-data professionals needed might increase from 5.7 million to 10.9 million.
How is AI defined? Europe states: AI is a machine's ability to exhibit human-like skills - such as reasoning, learning, planning and creativity. AI makes it possible for technical systems to perceive their environment, deal with these perceptions and solve problems to achieve a specific goal. The computer receives data - already prepared and collected via its own sensors, such as a camera - processes it and responds to it. AI systems are able to adjust their behaviour to some extent, analysing the effect of previous actions and working autonomously.
The Commission proposed that the EU invests in AI at least €1 billion per year from the Horizon Europe and Digital Europe programmes. The Recovery and Resilience Facility is also mentioned for providing a huge opportunity to modernise and invest in AI. Through all this the EU can become a global leader in the development and uptake of human-centric, trustworthy, secure and sustainable AI technologies.
However, when looking at the EU interventions and research activities chosen to address the transition challenge, there is also a reason for concern. Keywords in EU-calls are Cyber security, Cloud data & TEF, Quantum communication infrastructure, Deployment and Best Use of Digital Capacity and (governmental) Interoperability and EDIH infrastructure. Few interventions and calls are open for proposals for developing tacit solutions that address practical challenges, and therefore should focus more on the involvement of machine and robotic industries.
Nowadays digitisation not only seems to be becoming an end in itself, we are also spending more and more energy, capacity and time on it - while we were once promised that we would be given more space for what really matters in life. But the reverse is true. The very same thing is happening in food production, in which the digital transition is positioned as the key challenge to be addressed, as can be seen in the above mentioned section on the EU digital transition.
To the contrary, the real problems in sustainable food production are more concerned with:
Critique on current approaches to alleviate the above problems in food production can be summarised by: too much emphasis on data- and too much digital-centric. Data and digital technologies should be a means and not an end-goal on itself🙁. We are currently facing an enormous gap between decisions being advised and the actual production of food!
When looking at the various problems above one can directly see that NONE of those can be solved by digital technologies in a DIRECT way, simply because the digital world cannot act in the real world. Clearly there is something missing that digital technologies cannot deliver!
Engineering is essential for connecting the digital world to the real world. Research gives us understanding, but do we need to understand 100% before we start engineering? Can't we understand faster by engineering solutions?
Actual production (involving crop sowing, growth, crop maintenance, harvesting, post harvest processing etc) needs versatile handling. In order to address the mentioned problems intelligent autonomously operated machinery/robots are needed. Besides mechanical and electrical and software engineering, also engineering of artificial intelligence (AI) is required. These multiple disciplines allow for better optimal behaviours thanks to the extra available degrees of freedom.
AI has many faces. One part in this spectrum is data intensive AI with applications as image understanding (deep learning and more specifically transformer technologies) and recently enormous progress has been made in language understanding and generation (chatGPT). Its main field of science is informatics, and the associated skill is programming.
Another part within this same AI spectrum however, is Embodied AI [EMBOD-AI] with applications in autonomous machinery and robotics that allow for direct actuation in the real world. Its main field of science is control theory (cybernetics), and the associated skill is setting up control cycles. The latter presumes little knowledge compared to the first so consumes less effort.
This essay is based on my conviction that true intelligence can only be "Turing" tested by work done in the real world. This implies the involvement of all mentioned disciplines, and in particular mechanical and electrical engineering and that they all cooperate for acting in the physical world in a cyclic way.
My vision is that within 10 years the majority of the production work is done by many versatile humanoid robots, electrically powered, and positioned in smaller and more balanced settings.
The benefits are:
→ So, some nice challenges are lying ahead!
The required change of direction of (EU) agricultural policy makers all boils down to 2 changes:
1. Change of focus from data analysis towards actual production.
2. Application of Artificial Intelligence (AI) in achieving this.