The challenge of A.I. for Ag-tech and France's position in this domain (art. 1 of 4).

With the emergence of AI, Ag-tech is about to catch up with other sectors of the economy that have been on the path to industrialization for a long time, but French agriculture is slow to go onboard of the train to new technologies.

For many reasons mentioned in previous articles, to start with climate change, agriculture is a promising field of business. Beyond the path of agroecology, biotech and high-tech will be the driving force of a new green revolution. Ag-techs are for sure expected to play a major role in this segment of the primary sector which was resistant to industrialization. Booming technologies, which since the 2010s has resulted in the emergence of the cloud, the Internet of Things or Artificial Intelligence e.g., should reshuffle the cards. I believe A.I. is the breakthrough innovation susceptible to have a leverage over all the others, as its promise is to overcome the ultimate obstacle to understanding of life: the complexity inherent in its variability.

To unravel the links between living things, systemic thinking has until now conferred undisputed authority to all executives performing in this approach. In agriculture as well as in healthcare, it is the ability to think in this way based on a global approach that sets technical executives apart. It is also admited within the community of agronomists that aptitude for systemic thinking differentiates these engineers from their peers specialized in the hard sciences and therefore relying on logical-mathematical thinking. Technology is both a partner and a competitor for them and furthermore, they had to deal with computing since it started to the point that this industry - which has become their alias - became their main outlet. So far, computing has not yet dethroned the most talented leading specialist doctors who benefit from enviable prestige and privileges. For the same reason in greenhouse horticulture, crop managers with green fingers are well respected. I would add that in agronomy as in medicine, the experimental approach is key for dealing with doubt because the executive who works on living things knows he knows nothing. His personal talent places him at the heart of a trade which has until now resisted the standardization which goes together with systematization and ultimately, the use of digital technology.

With the generalization of AI to all economic activities, agriculture and medicine are about to catch up with other sectors, as evidenced by the spectacular boom in e-health and Ag-tech very soon.

AI is about to reshuffle the cards in greenhouses where their role will be essential. Today, it is the job of the crop manager to analyze the data produced by an increasing number of sensors. The Internet of Things makes it possible to multiply the measurement obtained by means of an increasingly assortment of sensors. Because new sensor systems provides information on the transpiration of the crop, its stem diameter or the strength of an electrical activity, whereas until now it was only intended to measuring the growth conditions in both the air and the rooting system. For total control over them, the crop manager must control a growing number of equipment furnishing an enclosure that is less and less open to the outside with the emergence of air con systems. LED lighting draw the border between horticulture and indoor farming of which vertical farming is a logical trend. The concept of Controlled Agriculture Environment (CEA) stands out as an alternative to the standard and mature model of the Venlo type multi-span greenhouse. For process regulation purpose, A.I. will take over from the arbitrary control of the crop manager to steer the changes in position of all actuators in the greenhouse, based on the many available data which are either measured or calculated.

In open field where the farmer has no control over climatic conditions and very limited control over irrigation as far as it mainly depends on the soil and the weather, the role of high tech is limited to precision agriculture which aims to adjust input and water supplies to the individual needs of each plant. This area of application of Ag-tech involves imaging and IoT which the power algorithms of Decision Support Tools that A.I. promises to make smarter. Under such conditions, the farmer just applies the recipe of a crop plan with little adaptation from one year over the next. From there, the acreage controlled by a farmer in field crops is limited only by his investment capacity and this is perhaps why, at about 70ha, the average size of our family farms is roughly the same as that of our foreign competitors.

In the most intensive farms which are soiless heated greenhouses, the financial weight of assets (the cost of an equipped greenhouse is commonly €2M/ha) is a limitation to the expansion but the main one remains the complexity of the occupation described above. It seems to me that controlling the greenhouse based on the big amout of data generated by the sensoring system limits the acreage controlable by the grower to approximately 4ha. The actual figure for a family farms in France for such greenhouses is currently 4.5ha in France while in the Netherlands where horticultural companies have for long salaried crop managers, it exceeds 6ha with a standard deviation likely higher.

For the same reasons as in high end greenhouses, I think that scales savings do not work in the traditional seed business where mid-sized companies - agile and adaptable - are succesful whereas in agrochemicals, systematization has made possible extreme concentration of the firms in this sector. Just like the crop manager in a greenhouse, the gifted seed breeder with an intuition built upon a long lasting experience, is key in the seeds. To unravel the links between the genetic code and the expression of genes, A.I. promises to give meaning to the mass of data provided by genotyping and sequencing tools which have arisen in recent years with the invention of PCR especially. In the 2000s, these new research means added to marker-assisted backcrossing enabled seed companies stepping forward, after they have been for many years simple farmers specializing in the custom production of seeds at the era of mass selection of population varieties. I guess with the availability of AI-based software packages, the current time will be one of an even greater technological leap for seed companies.

As an observer of the horticultural sector where penetration of AG-tech is potentially the highest, I note however that France remains behind while it is still at the head of the powers European agriculture.

Proof of this delay of French agriculture in technology is that with 700 members or so, the Agro-Digital group of the UNIAGRO association, which brings together former agronomic and agricultural engineers, is mainly made up of curious retirees or IT executive working outside agriculture, while engineers in Ag-tech represent less than 5% of the group.

In the publications it administers, the Academy of Agriculture, which is at last opening itself to the world (the articles available on its portal have finally been translated into English since the beginning of the year), gives the main place to life sciences but it reserves the smallest part for Ag-tech with again, less than 5% of publications dedicated to this subject.

INRAE, whose explicit vocation is Agricultural and not technological research, doesn't have any skilled researcher in high tech. The same applies to the CTIFL (Technical Institute in charge of Fruits and Vegetables) while technology is about to play a major role in the controlled production of veggies in greenhouses.

At the Dutch world reference show GREENTECH dedicated to intensive horticulture, France is only represented by a few visitors and exhibitors while it organizes its own domestic show dedicated to so-called specialized crops in Angers: SIVAL. At around 25,000 visitors, the attendance of the two events is the same but our national show gives pride of place to inputs and machinery while the fair in Amsterdam stands out for the supremacy of Ag-tech. In France, technology is th topic of a couple of regional events on e.g. robotics, connected agriculture or photovoltaic greenhouses. On the other hand, with a number of visitors of the same order as SIVAL and GREENTECH, the annual TECH&BIO fair which is held each year in Valence, confirms our preference for life sciences.

France is also invisible in the press specialized in Ag-tech applied to horticulture since the leading daily global newsletter HORTIDAILY refers to France less than 10% of issues. Each issue having at least 30 articles, this means that France's appears in less than 1 article in 300. These articles highlight sometimes a French grower, a successful fundraising by an AG-tech startup but more frequently, communicate on supermarket chains. This is a sign that on the international horticulture market, France is seen above all as an outlet while its position remains anecdotal in technology. To date, France is still not in the race for the autonomous greenhouse that this publication periodically reports on.

We can understand the choice of HORTIDAILY since in horticulture, French farms are all the same as their degree of intensification increases with penetration of new technologies. While there is a great diversity of farms in open field, the production tool of high end growers is identical and designed using mature and proven technologies. City farming remains low-tech in France and within the C.E.A., there is hardly any indoor farm while aquaponics or algae cultivation exist in the form of small, generally low-tech operations.

Most of French Ag-tech technology suppliers are startups positioned in trendy markets such as precision agriculture, robotics, LED lighting, aquaponics or indoor agriculture, among other examples. Gathered under the banner of FERME DIGITALE (Digital Farm), these 120 partly subsidized startups show on at our national agricultural show but their influence does not exceed the limits of our borders and those which export or have branch offices abroad, are the exception that proves the rule. The major players, some of whom are approaching the unicorn stage, are rarely French and in most cases "pure players" of global stature who can possibly rely on the financial strength of other “cash cow” markets to finance their R&D.

Following this first article showing from facts and figures that our agricultural sector is living science driven but very slow to jump on the high-tech train, I will soon release a second article online where I will formulate hypotheses on the reason why of this delay. From there, I will discuss the consequences before proposing solutions for catching up with other sectors of the economy and one of them will surprise you.

#agritech #agtech #Ag-tech #AI #agriculture #horticulture #indoor #city farming