The easiest way for an organisation to lose itself is to introduce AI without clarity about what it may and may not be used for.<\/p>\n
That might sound obvious, but practice tells a different story. Purpose is rarely made explicit before an AI system goes live. Not because nobody cares about it, but because purpose feels abstract compared with a working demo, and abstract conversations often lose out to concrete deadlines.<\/p>\n
Without a sharp sense of purpose, AI quietly infiltrates your culture.<\/strong><\/p>\n Technology is never neutral. That is not a philosophical statement, it is a practical warning. Every AI system carries a view of people, baked into the choices made during design, even when nobody has consciously named those choices as a view of humanity. Are people in that system a cost to be minimised, or a source of meaning? Are mistakes risks to be prevented, or raw material to learn from? Is deviating from the norm undesirable behaviour, or precisely the mark of professionalism?<\/p>\n Those questions are usually not asked before a system is procured. They only become visible afterwards, in how the system actually behaves, and by that point it is already baked into processes that are no longer easy to reverse.<\/p>\n There are three decisions about purpose that I believe every organisation should make, preferably before the first pilot goes live, not after.<\/p>\n The first decision concerns what AI may contribute to. Tie that contribution explicitly to strategic goals: better service, shorter turnaround times, less administrative burden. But be at least as explicit about what AI never replaces. Not the final moral judgement. Not the conversation with employees. Not the relational core with customers or residents. A sentence I often recommend organisations write down literally is this: AI supports the work of professionals; it never makes final decisions about people without human judgement. That sentence sounds simple. It becomes binding the moment you actually use it as a test against every new application.<\/p>\n An organisation that does not choose explicitly chooses implicitly. And the implicit choice is almost always the one that meets least resistance in the moment, not the one that best honours what the organisation genuinely wants to be.<\/strong><\/p>\n The second decision concerns values that must never come under pressure. Choose three to five values, no more, because a long list of values never becomes a workable instrument. Think of human dignity, fairness, reliability, transparency, inclusion. Then translate each value into a concrete design principle. The right to explanation, for instance: anyone assessed by a system must be able to ask why, and must receive an understandable answer. Room for dissent without repercussions: nobody should suffer a disadvantage for questioning an AI outcome. Logging human override: when someone deviates from the system, record it, not to monitor that person, but to allow the organisation to learn from the patterns that emerge.<\/p>\n The third decision concerns red lines, the things you explicitly do not do. No hidden performance profiling without people knowing it happens. No automatic refusals in vulnerable cases without a human with an opposing view having looked at it. No emotion monitoring without a compelling reason and without the consent of the works council. These red lines might feel like restrictions on what is technically possible. They are. That is precisely their function.<\/p>\n It is worth dwelling on this explicitly, because red lines are often interpreted as distrust of technology, when in reality they are something else: a translation of what the organisation already knew about its own values, before AI entered the picture. A red line around hidden profiling existed implicitly long before there was an algorithm capable of carrying it out. What AI does is test that implicit boundary, simply because it suddenly becomes technically possible to cross it without anyone immediately noticing. Making the red line explicit is therefore not a new restriction. It is saying out loud something that already applied, at the moment it matters.<\/p>\n People listen less to words than to practices.<\/strong><\/p>\n This point deserves repeating, because it strikes at the heart of why purpose without practice is worthless. If your first pilot revolves around monitoring, the organisation receives a message that was never spoken aloud but comes through loud and clear: we trust the system more than you. If you deploy AI primarily to relieve people and strengthen their expertise, the organisation receives a different message: we take your judgement seriously.<\/p>\n That same mechanism, incidentally, does not only operate at the level of the organisation as a whole, but also at the level of individual managers. A manager who says they trust their team’s judgement, but double-checks every decision with an AI system before agreeing to it, communicates something different from what their words claim. That need not be a conscious contradiction. It is often simply unconscious behaviour arising from uncertainty about a new instrument. But the effect on the team is the same, regardless of the intention behind it.<\/p>\n How do you approach this practically, within a timeframe a busy organisation can actually afford? It can be done in two hours, if you organise it sharply. Organise a session with the management team around three questions: what may AI work on for us, which values are non-negotiable, and what are our red lines? Capture the outcome in a one-page compass, supplemented by a second page full of concrete example situations: this we do, this we do not. And place that compass alongside your existing vision of leadership and culture. Does the story it tells match the story you want to show within the organisation?<\/p>\n What you can do today is formulate your own purpose for AI in one sentence, in plain language, without jargon. What you can do over a slightly longer horizon is test three existing practical cases against the values you have established, to see whether they actually hold up or only sound good on paper. And what is best avoided is what I call value-washing: posters of values on the wall that never once led to a concrete design choice.<\/p>\n Can you, today, credibly answer the question of what AI may be used for at your organisation, and what it absolutely may not?<\/p>\n If the answer is no, that is no cause for panic. It is an indication of where the work now needs to begin.<\/p>\n Notes for those who wish to read further:<\/em><\/p>\n The easiest way for an organisation to lose itself is to introduce AI without clarity on what it may and may not be used for. Part two of a series on implementing AI without losing yourself.<\/p>","protected":false},"author":1,"featured_media":926,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[26,81],"tags":[89,100],"class_list":["post-5016","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog","category-english","tag-english","tag-leadership"],"_links":{"self":[{"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/posts\/5016","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/comments?post=5016"}],"version-history":[{"count":1,"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/posts\/5016\/revisions"}],"predecessor-version":[{"id":5125,"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/posts\/5016\/revisions\/5125"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/media\/926"}],"wp:attachment":[{"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/media?parent=5016"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/categories?post=5016"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dbvp.nl\/en\/wp-json\/wp\/v2\/tags?post=5016"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}\n