What Holds When Every­thing Vibrates

A few days ago, we stood in an indus­tri­al hall with thir­ty thought lead­ers and shapers of the future. In front of us, on a small table, lay a wood­en bowl with five stones we had gath­ered only days before in Nor­way, beyond the Arc­tic Cir­cle. Behind us lay two dense half-days of AGORA, the for­mat Dr. Bern­hard von Mutius con­venes once a year, an after­noon and a morn­ing filled with con­ver­sa­tions about peo­ple, robot­ics, AI, and about the ques­tion of where the future of our indus­try actu­al­ly lies. On the after­noon of the sec­ond day, when the cir­cle had already shared much togeth­er, our turn came.

No one had quite expect­ed the title of our con­tri­bu­tion: „Sta­bilis­ers in AI dis­rup­tion – what holds when every­thing vibrates?” You are warm­ly invit­ed to read along and think with us about why we believe this ques­tion can­not be post­poned.

The wrong reac­tion: oper­a­tional fren­zy

You prob­a­bly know this from your own organ­i­sa­tion. AI is accel­er­at­ing, the pace is ris­ing, and the reflex is almost every­where the same – more speed, more tools, more projects, more con­trol, more of every­thing we already know. We call this the pres­sure cook­er of oper­a­tional fren­zy: AI projects are set up, pilots launched, work­shops organ­ised, and much of it is well-inten­tioned, yet remains on the sur­face of what is actu­al­ly mov­ing beneath.

Some time ago we came very close to an ice­berg in Green­land, in a kayak, cir­cling it care­ful­ly and being warned repeat­ed­ly not to come too close, when sud­den­ly the ice­berg turned over before our eyes: the peak we had just been admir­ing sank into the icy sea, and the pre­vi­ous­ly hid­den under­side faced the sun for the first time. It was an image that has stayed with us ever since, because it shows some­thing we keep encoun­ter­ing in organ­i­sa­tions as well.

What is vis­i­ble – the process­es, the struc­tures, the tools – is only the peak, and the actu­al dynam­ics lie beneath, in atti­tudes, in images of pow­er, in rela­tion­ships. AI does not mere­ly ampli­fy what is already there but also shifts the con­di­tions under which struc­tures can hold, so that even struc­tures func­tion­ing well today may come under pres­sure tomor­row. Nora Diet­rich, an expert on men­tal health in the work­place, speaks in this con­text of organ­i­sa­tion­al burnout – organ­i­sa­tions exhaust­ing them­selves under the weight of their own ini­tia­tives – and we would add that AI ini­tia­tives with­out organ­i­sa­tion­al matu­ri­ty are the surest path there.

Five stones, five sta­bilis­ers

We devot­ed the open­ing of our con­tri­bu­tion at AGORA to the metaphor of our stones. We had only just gath­ered them at the Arc­tic Cir­cle, in a place where the effects of cli­mate change are very strong­ly felt: glac­i­ers are melt­ing, coast­lines are shift­ing, the land itself is trans­form­ing before our eyes. And yet these stones are what they have always been – stone, car­ry­ing change with­in them­selves with­out being dis­solved by it. This is pre­cise­ly what we mean when we speak of sta­bilis­ers: not the pre­ven­tion of change, but the capac­i­ty to car­ry it.

We took the stones out of the bowl one by one and placed them into the room – one for each sta­bilis­er, one after the oth­er, vis­i­ble in the mid­dle, where there had been noth­ing before. Per­haps, as you read, you might place your own stones in your mind in a spot where there had been noth­ing before.

Sta­bilis­er 1: Dis­trib­uted respon­si­bil­i­ty

AI decen­tralis­es knowl­edge, and from this it fol­lows almost nec­es­sar­i­ly that respon­si­bil­i­ty, too, must be decen­tralised – not as a nice con­cept, but very con­crete­ly as deci­sions made where com­pe­tence and con­text meet. We work with the image of „dial, not switch”, because this is not a bina­ry choice between human and machine, between cen­tralised and decen­tralised, but a mat­ter of sit­u­a­tion­al adjust­ment: some deci­sions the AI makes bet­ter, some the human does, and most emerge best when both work togeth­er in clear­ly dis­trib­uted roles.

Respon­si­bil­i­ty in this under­stand­ing is not sim­ply hand­ed out, it is nego­ti­at­ed, and it is por­tioned in the size that some­one can actu­al­ly car­ry, because what is too large pro­duces over­whelm, and what is too small pro­duces mere execu­tors – and nei­ther of these will serve us well in the AI era.

Sta­bilis­er 2: Dia­logue and col­lec­tive intel­li­gence

AI can help us find good ques­tions, some­times even bril­liant ones, but it can­not sense which ques­tion, in this room, with these peo­ple, at this moment, cre­ates move­ment, and which mere­ly cre­ates busy­ness. The greater dan­ger of AI, then, seems to us not that it will take our work away, but that we will for­get how to talk to each oth­er – that we will trust the algo­rithm more than our col­league and ignore what is left unsaid, because the data, after all, appears to be right.

In every organ­i­sa­tion there are sen­tences no one con­tra­dicts, not because every­one agrees, but because silence is more com­fort­able; and silence in organ­i­sa­tions is nev­er absten­tion, but a form of con­fir­ma­tion. The temp­ta­tion in the AI era is now that we will increas­ing­ly hide behind data instead of say­ing what is uncom­fort­able in the room, and it is pre­cise­ly this temp­ta­tion that must be con­scious­ly resist­ed.

Sta­bilis­er 3: Con­scious (un)learning

For us, this is the cen­tral sta­bilis­er of the AI era, and it has two sides that belong togeth­er. One side is learn­ing: build­ing new capa­bil­i­ties, and doing so for every­one, not only for the IT depart­ment, because in the AI era learn­ing time must nat­u­ral­ly be work­ing time. The oth­er side is hard­er and remains large­ly unad­dressed in many organ­i­sa­tions – unlearn­ing, the con­scious let­ting go of old assump­tions that no longer hold but that no one dares to name.

We mean, very con­crete­ly, those sen­tences that exist side by side in our organ­i­sa­tions every day, with­out any­one dar­ing to point out the con­tra­dic­tion:

„Col­lab­o­ra­tion is our most impor­tant asset” – yet we still reward indi­vid­ual per­for­mance.

„We want a learn­ing cul­ture” – yet any­one who makes a mis­take faces a career prob­lem.

„We are built on trust” – yet every deci­sion requires three sign-offs.

These are not small incon­sis­ten­cies but sys­temic con­tra­dic­tions, and AI makes them mer­ci­less­ly vis­i­ble, because every AI ini­tia­tive will fail pre­cise­ly where the sys­tem speaks a dif­fer­ent lan­guage than the strat­e­gy. Rik Vera, a Bel­gian futur­ist, puts it sharply when he observes that our board­rooms are cal­i­brat­ed for answers and not for ques­tions, that KPIs do not mea­sure uncer­tain­ty and share­hold­ers do not applaud unlearn­ing – and that, pre­cise­ly for this rea­son, we out­source our curios­i­ty.

Sta­bilis­er 4: Rad­i­cal kind­ness

That we speak of kind­ness in a busi­ness con­text may come as a sur­prise, and we want to clear up a mis­un­der­stand­ing from the out­set: we do not mean nice­ness, nor har­mo­ny at any cost, but a con­scious choice for a respect­ful, atten­tive way of deal­ing with one anoth­er – espe­cial­ly under pres­sure, espe­cial­ly in dif­fer­ence, espe­cial­ly in con­flict. Kind­ness in this sense is not a feel­ing that arrives when the sit­u­a­tion is favourable, but a stance that one choos­es pre­cise­ly when the sit­u­a­tion does not invite it.

In times of tech­no­log­i­cal over­load, peo­ple need psy­cho­log­i­cal safe­ty, because no one who is afraid will learn, no one who is afraid will exper­i­ment, and peo­ple who are afraid will hide behind rou­tines instead of open­ing them­selves to what is new. Kind­ness, there­fore, is not a soft top­ic but the social lubri­cant that pre­vents AI change from turn­ing into resis­tance, and emo­tion­al con­ta­gion works in both direc­tions: fear of AI spreads just as read­i­ly as curios­i­ty, and as the per­son respon­si­ble we have a say in what we sow.

Sta­bilis­er 5: The art of bal­ance

In our gar­den we have set up a slack­line, and any­one who stands on it once learns very quick­ly that bal­ance does not come from stand­ing still but from con­stant, fine cor­rec­tions – and who­ev­er over­cor­rects ampli­fies the swing and falls. The same is true with AI: there is the adven­ture trap, bet­ting every­thing on AI with­out a foun­da­tion, and there is the safe­ty trap, ignor­ing AI in the hope that it will pass, and the art lies in between, in a sta­bil­i­ty that serves as a spring­board, because those who do not know whether the ground will hold do not jump.

What strikes us as par­tic­u­lar­ly uncom­fort­able in the AI era is the sug­ges­tion of max­i­mum effi­cien­cy – every posi­tion filled once, every process trimmed to the bone, every buffer opti­mised away – which works as long as every­thing goes accord­ing to plan, and the moment some­thing unfore­seen hap­pens, and it always does, the reserves are gone. Redun­dan­cy, para­dox­i­cal as it sounds, is there­fore not waste but the pre­con­di­tion for peo­ple to dare to jump into what is new in the first place.

What Europe has – and what you can do now

At the end of our talk we point­ed to the five stones in the cen­tre of the room: vis­i­ble, tan­gi­ble, hold­ing, while out­side every­thing was vibrat­ing.

Europe has some­thing Sil­i­con Val­ley does not have, and it is worth say­ing with­out false mod­esty: we have a deep organ­i­sa­tion­al sub­stance, a cul­ture of rela­tion­ships that has grown over gen­er­a­tions, and an entre­pre­neur­ship that does not look for a quick exit but for the next gen­er­a­tion. This is not a lag but a foun­da­tion – and it is pre­cise­ly this foun­da­tion that we need in order to inte­grate AI into our organ­i­sa­tions in a way that is respon­si­ble, human, and sus­tain­able.

Three invi­ta­tions we would like to leave with you, the same ones with which we closed our talk.

The first is a ques­tion you can ask in your organ­i­sa­tion, open­ly and togeth­er: what do we need to unlearn so that AI can actu­al­ly become use­ful, and which old assump­tions are hold­ing us back?

The sec­ond is an invest­ment deci­sion: invest in human infra­struc­ture, not as a coun­ter­weight to tech­nol­o­gy, but as a pre­con­di­tion for it. Because trust, dia­logue, and dis­trib­uted respon­si­bil­i­ty are not costs, but the invest­ment in a true com­pet­i­tive advan­tage.

And the third is a stance: open up, with­out fear, with curios­i­ty, with a sense of new begin­nings.

Our five stones we left in the room after the talk, and they are now some­where out in the world. The ques­tion we placed at the end, how­ev­er, belongs to you as well:

If AI took over thir­ty per­cent of the tasks in your organ­i­sa­tion tomor­row – what would you do with the freed ener­gy? And: would your organ­i­sa­tion even allow it?

Tell us what you think. We look for­ward to the con­ver­sa­tion.

With thanks to Dr. Bern­hard von Mutius and the AGORA cir­cle for two days full of sub­stance, and to Peter Edel­mann for the hos­pi­tal­i­ty.