The Good and the Bad News About AI

The bad news is that everyone is currently waffling on about AI. The good news is that this is completely normal.

Not good normal. Not elegant normal. Certainly not, “Here is a sector calmly working through a major capability shift with mature governance and a shared understanding of risk” normal. Let’s not get carried away. It is normal because this is what humans do whenever technology changes the shape of work.

We panic, posture, overclaim and underthink. We appoint gurus, invent jargon and confuse confidence with competence. Before long, someone is running a webinar called something like Unlocking the Future of Transformational AI Synergy, and we have all lost the will to live by slide four.

Then something interesting happens. While everyone is still arguing about whether the technology is good or bad, it quietly becomes infrastructure. It becomes ordinary, and ordinary is usually the point at which the real transformation begins.

That is what I think is missing from much of the current AI conversation. The interface feels new. The speed feels new. The level of access feels new. The scale and drama certainly feel bigger. But the pattern underneath it is not new at all.

We have been here before.

We did this with websites, email, search engines, content management systems, social media, cloud storage, smartphones, remote working, Zoom, Canva, CRMs, dashboards, automation, online booking systems and shared drives. Each one was introduced as a game-changing platform that would revolutionise everything. Most eventually became another tab we keep open while trying to remember which variation of our password we used.

AI is undoubtedly a major change. It is powerful, fast, uneven and full of risk. It is already changing how people work, whether their organisations have written a policy or not. But it is still change, and humans have always been rather strange about change.

We love innovation when it appears in a keynote, a funding bid, a strategy document or a conference panel where everyone says “ecosystem” seventeen times and nods as if they have personally invented the future. Actual change is different.

Actual change asks people to alter a process, learn something new, admit that they do not understand a tool, rethink parts of their job, examine their data, update their governance and change their access permissions. It may even ask them to stop storing important business information across six inboxes, three spreadsheets, a WhatsApp thread and inside their co-worker Karen’s head.

Naturally, everyone hates that bit.

That is, in many ways, where we are with AI. It has not suddenly created all the problems businesses are now panicking about; it has exposed them. It has shown us how many organisations do not know where their data is, how many teams lack clear processes and how much work has never been properly mapped, measured, documented or understood.

It has also exposed how much professional knowledge is informal, invisible or trapped in people’s heads. Information is scattered across tools purchased by different departments, in different years, by different people, some of whom have since left.

AI has wandered in, cheerfully offering to automate things, and everyone is panicking because it turns out nobody can quite explain what the “things” are.

That is not AI’s fault. AI is holding up a mirror, and some people are blaming the mirror for the face.

The tool is never the whole story

The useful part of this disruption is that once technology exposes the cracks, we have an opportunity to fix them.

That is what gets lost amid the drama. People keep asking whether AI is good or bad, but that is not a particularly serious question. A hammer is useful when you are building a shelf and rather less welcome when someone is smashing a window. A spreadsheet is useful when the person using it understands the data, but less so when it contains three hidden tabs, six broken formulas and the only copy of the company forecast. Canva can help someone communicate clearly, or it can give a person with no design sense the confidence to use twelve fonts and a gradient.

The tool is never the whole story. The user matters. The context matters. The data, process, governance and quality of the thinking all matter.

The AI debate has become exhausting because people keep trying to turn the tool into a moral personality. They want it to be either the villain or the saviour, magic or theft, genius or nonsense. It must apparently be either the end of human work or the beginning of a glorious productivity age in which everyone gets Friday afternoons back to paint watercolours and think strategically.

The truth is more irritating: AI is a tool.

It is a very powerful tool that can do remarkable things in the right hands, with the right inputs and within the right boundaries. It can also produce absolute sludge when it is used badly, lazily or vaguely, particularly by someone who thinks “write me a viral post” constitutes a strategy.

But it is still a tool, and tools eventually get absorbed.

At first, everyone performs their relationship with the new thing. Some people become early adopters, while others become sceptics, evangelists or doomers. Some become experts three working days after creating an account. Others become thought leaders, which is often what happens when a person has not yet found a useful job for their confidence.

Eventually, the fuss settles. The terrible uses remain terrible, the useful uses become normal, and serious people get on with serious work. In time, everyone forgets they were ever confused.

We have already lived through this with websites

The history of websites gives us a useful comparison.

Early business websites were essentially printed brochures that someone had emotionally blackmailed into becoming HTML. Businesses had them because they had been told they needed one, although few people knew what should go on them. They were hard-coded, clunky, slow and frequently opened with “Welcome to our website,” accompanied by a stock photograph of a handshake or a woman laughing alone with a salad.

Then people became frustrated because they could not update them, so we built content management systems. When those systems proved too complicated, we developed simpler platforms, page builders, themes, templates, drag-and-drop editors, plugins, integrations, training guides, onboarding processes and help bubbles that asked, “Would you like assistance?” while the user screamed internally.

Even drag-and-drop tools still required decisions, however. They required structure, taste, judgement, planning, content and a basic willingness to read the instructions. In response, the industry built more rails, frameworks, templates and automation.

For decades, the feedback from users was some version of: “I cannot do this. It is too complicated. Can the software not simply do more of it for me?”

Now the software can do a startling amount of the work people spent years saying they could not do, and everyone is furious.

That is, if nothing else, quite on brand for humanity.

The technology industry has spent decades removing friction, hiding complexity, simplifying interfaces and building tools that allow more people to achieve things without understanding every technical layer underneath them. AI is the next logical stage in that process, yet people are reacting as if the toaster has started quoting Nietzsche.

Plenty has changed. The speed, accessibility, scale and risk profile are different. The ability to produce plausible-looking nonsense at volume has certainly increased, and LinkedIn may never recover. But the underlying story has not changed as much as people think.

Humans remain at the centre of it.

Some AI systems can take a rough idea and structure, test, compare, translate, summarise or reframe it faster than many organisations can locate the latest version of their own logo. Some can work across processes in ways that would have felt like witchcraft ten years ago.

Others are shockingly bad: truly, magnificently, “Who allowed this near a business process?” bad. The people least likely to recognise this are often those using it with the confidence of a man in 2008 insisting that his nephew can build the company website because he is “good with computers.”

Again, none of this is new. Bad tools, users, strategies and data have always existed. Bad procurement has practically had its own protected habitat in the UK for decades.

AI has not invented poor decision-making. It has given it a faster horse.

The bad news is that we are having the wrong arguments

The real bad news is not that AI exists. It is that people are managing to have almost all the wrong arguments about it.

We have people debating whether AI is good or bad as if they are reviewing a lasagne. Some insist that everything it touches is soulless while spending their entire working day using digital systems whose underlying mechanisms they never question. Others publish AI-generated content warning us about the dangers of AI-generated content, which is almost beautiful in its lack of self-awareness.

Meanwhile, people sell “AI mastery” after three afternoons and a Canva carousel, while businesses panic about robots taking jobs even though half their operational knowledge lives inside an inbox labelled “misc.”

Organisations with no data policy, no recent access review, no practical staff training and no shared understanding of which tools employees are permitted to use are suddenly asking whether AI is safe.

Compared with what?

Safe compared with the spreadsheet everyone downloads and emails around? Safe compared with customer data stored in someone’s personal Gmail because it was easier? Safe compared with the intern who has access to everything because nobody configured the permissions correctly? Safe compared with the board pack sent to thirteen people, three of whom left last year?

What about the CRM nobody trusts, the Google Drive nobody has structured, the website nobody owns or the social media passwords kept inside a note called “passwords”?

This is why some of the panic feels so bizarre. People are terrified of the Terminator while Derek from accounts is pasting sensitive client information into a free browser tool because it gave him a really good summary.

The robot uprising is not currently my biggest concern. Derek is much higher on the list.

That does not mean AI risk is imaginary. There are serious questions concerning data protection, copyright, bias, misinformation, security, intellectual property, procurement, accountability, workforce impact and quality control. A business using AI without considering those issues is not being innovative. It is being reckless with better branding.

But the answer is not panic. It is literacy, policy, judgement and governance.

Organisations need to understand what the tool is, what it is not, where the data goes, who remains responsible, what needs checking and what should never be uploaded. Most importantly, they need to know what the business is actually trying to improve.

Unfortunately, that is harder than buying a subscription.

Buying a tool requires a card, a login and a slight sense of destiny. Changing how people think, work, decide, check and take responsibility is slower, less glamorous and much harder to fit into a post titled “Five Ways AI Will Transform Your Business by Friday.”

Yet that is where the real work is, and it always has been.

Digital transformation was never just about software

Software is the visible part of digital transformation. Beneath it sit people, incentives, confidence, fear, habits, workflows, leadership, culture, data and accountability. There is also the alarming possibility that nobody truly understands how a process works until the person who normally manages it goes on holiday.

AI has made those realities harder to ignore.

It walks into the room and asks, “What would you like me to automate?” Half the room falls silent because the answer is apparently, “A collection of undocumented workarounds held together by goodwill, memory and three spreadsheets called final, final-final and actual-final.”

Again, not AI’s fault.

This is why I find the emotional relationships people are developing with AI unhelpful. I do not “like” AI, but I do not dislike it either. I do not gaze lovingly at a spreadsheet, ask Canva how it is feeling or wonder whether Google Analytics respects me as a person.

It is tooling.

We should use it well, carefully and critically, with proper data boundaries, human review and a clear understanding of what good looks like. Ideally, we should avoid uploading the company, client list, board papers, commercial strategy and confidential meeting notes into something whose terms nobody has read.

That would be lovely.

AI has automated mediocrity, not invented it

The quality problem is real. There is now a great deal of AI-generated content, and some of it is dreadful. Not because AI touched it, but because nobody with judgement touched it afterwards.

It has a strange laminated quality. Everything sounds professional, yet nothing says anything. It is fluent, polished and structurally competent, while remaining completely empty. It frequently begins with “In today’s fast-paced digital landscape,” with the energy of a machine trained entirely on airport business books and procurement portals.

However, bad content existed before AI. So did bad websites, weak strategies and pointless social posts. AI did not invent mediocrity; it automated its production. That is irritating, but we should not pretend a golden age of human communication was interrupted only last Tuesday.

The relevant question is not whether AI was involved. The question is whether the result is any good.

Does it say something? Is it true, useful, specific and appropriate? Does it understand the audience? Has somebody checked it? Has anyone contributed judgement, experience, taste, evidence, clarity or an actual point?

That is how the work should be judged, rather than turning every social post into a witch trial in which people sniff the punctuation for signs of automation.

Some human-written content is terrible. Some AI-assisted content is extremely useful. Some AI-generated content is bland nonsense. Some entirely human content is also bland nonsense, except it took longer and involved more ego.

Quality is not created by the absence of tools. Quality is created by judgement.

More output is not the same as more value

This distinction matters because businesses are increasingly tempted to confuse output with value.

AI can help produce more posts, drafts, documents, summaries, proposals, plans, ideas and content. But more is not automatically better. Frequently, it is simply more mess wearing a productivity badge.

If the thinking is weak, AI can help scale weak thinking. If a process is broken, it can automate the broken process. If the data is poor, it can produce confident answers from unreliable foundations. If the brief is vague, it can deliver a beautifully formatted version of the vagueness.

That is not transformation. It is faster chaos.

We should probably stop calling something innovative every time someone attaches a chatbot to a bad process.

A more serious question is where business data is going, who has access, what staff are uploading, how outputs are reviewed and whether client information is protected. Organisations should know which tools are approved and whether anybody can explain the difference between using AI to support thinking and allowing it to make decisions nobody understands.

People will argue passionately about the soul of creativity and then upload a commercially sensitive document to a free tool with the same energy as someone throwing receipts into a carrier bag.

This is where we need to grow up.

AI is not separate from data protection, governance, cyber security, customer trust, business process, staff development or operational risk. It sits across all of them, pulling on existing weaknesses and exposing them.

Treating it as a novelty is therefore dangerous. It is no longer simply a fun writing tool, a productivity hack or something the marketing team can experiment with while everyone else carries on as before. It is already influencing how people search, write, plan, compare, summarise, decide, learn and communicate.

The adult conversation is not, “Do we like AI?” It is, “Where is it already being used, by whom, for what purpose, with what data, under which rules, with what checking, and what happens if it is wrong?”

Less exciting, perhaps, and much harder to sell through a carousel, but considerably more useful.

AI is the latest mirror

It is understandable that everyone is tired. Politics is grim, the internet is grim and the economy is grim. After years of disruption, the general mood sits somewhere between “deeply over it” and “one more login and I will walk into the sea.”

But it is not all ChatGPT’s fault.

AI did not invent weak leadership, poor communication, bad procurement, boring content, digital exclusion or people pretending to understand things in meetings. It is simply the latest mirror.

What we are seeing reflected is the quality of our thinking, data, processes and decision-making. That is uncomfortable, but it is also useful.

When AI exposes unclear processes, messy data, undertrained staff, vague governance, disconnected systems and a digital strategy consisting mainly of subscriptions with different invoices, the problem is not that the technology has become too clever. The problem is that the foundations need attention.

That is not a reason to despair. It is a reason to stop having silly arguments.

AI will not save a business that does not know what it is trying to do. It will not repair weak strategy, create good judgement, turn poor data into reliable intelligence or make a vague offer compelling. It will not heal a broken process, create confidence in an unsupported team or make leaders accountable when they are determined to hide behind novelty.

It can, however, help capable people do certain things faster, better or differently. It can support drafting, structuring, comparing, summarising, reviewing, testing and learning. It can reduce blank-page anxiety, make complex information more accessible and help small teams punch above their weight.

It can help people think through options, provided they understand how to interrogate the response.

But it cannot care. It cannot take responsibility. It cannot understand an undocumented business better than the people who run it. It cannot make a good decision in a context nobody has explained. It cannot replace the human work of knowing what matters.

That remains our responsibility.

The good news is that people are more ready than we think

Although people may not use official terminology or describe themselves as digitally confident, they have already absorbed an extraordinary amount of technological change.

They know how to search, compare, message, share, upload, download, review, book, pay, post, edit, comment, screenshot, forward, scan, swipe, stream and subscribe. They know when a website feels wrong, when an app is easier than a form and when a process is wasting their life. They recognise when technology genuinely helps and when it merely creates more work with a shinier logo.

That is not nothing. It is the foundation.

Business owners who once needed a developer to change a phone number can now edit their own pages. People who could not crop an image now produce social graphics in Canva. Teams that once passed documents around as email attachments now collaborate through shared drives. People who printed route directions now become irritated when Google Maps takes three seconds to reroute them. Those who once distrusted online banking now move money from the sofa without a second thought.

We have become incredibly digitally capable compared with where we started. We often fail to recognise this because the benchmark keeps moving.

The population has moved with the technology—not evenly, perfectly or without exclusion, confusion and risk, but significantly. People who once dismissed social media as something for teenagers now run businesses through Facebook groups, LinkedIn posts, Instagram reels, WhatsApp communities and TikTok shops. Those who thought cloud storage sounded suspicious now panic when a document is not instantly available on their phone. People who refused video calls now have a favourite Zoom angle.

The people changed. The work changed. The expectations, tools and baseline changed.

AI is moving the baseline again. That is uncomfortable, but it is not unprecedented.

AI is the next step in removing friction

In many ways, AI is simply the latest stage in a long story of asking technology to remove friction.

The request has remained remarkably consistent: make it easier, faster, editable, searchable, shareable, visual, integrated and automated. Help me avoid starting from scratch. Allow me to achieve the result without understanding every technical layer underneath it.

The technology industry, in its chaotic, flawed, brilliant and occasionally cursed way, has spent decades answering that request.

We developed content management systems so people did not have to code every page, templates so they did not have to design from scratch, and search engines so they did not have to remember where information lived. We built CRMs so customer relationships did not exist entirely inside one person’s head, automation so people did not repeat the same tasks endlessly, and integrations so information could move between tools.

We created analytics so businesses did not have to guess what was happening and dashboards so they could ignore those instead of ignoring spreadsheets.

Now we have tools that can draft, summarise, translate, analyse, sort, compare, explain and assist.

That is not an alien invasion. It is product development.

It is the logical outcome of thirty years of asking, “Can the system not do more of this for me?” Now, increasingly, it can.

Not perfectly, safely by default or without human judgement. Not without data literacy, policy, risk management and governance. But it can.

Perhaps the appropriate response is therefore not mass hysteria, but a cup of tea, a sensible access policy, some staff training, a data audit and a little emotional regulation.

The panic is not helping, but neither is the worship. Both are lazy in different outfits.

Doomers talk as though every AI system is a sentient death machine waiting to destroy the arts, education, democracy and your nan. Evangelists speak as though it will solve productivity, loneliness, social care, climate change, procurement, administration and the printer by next Thursday.

Neither position is serious enough.

The serious position is more boring and far more useful. Some AI is excellent, some is rubbish, some is dangerous in the wrong context and some is genuinely transformative in the right one. Most people need greater literacy. Most businesses need better data hygiene. Most organisations need clearer rules, while teams need permission to experiment safely.

Leaders must understand enough to ask intelligent questions, and users need to stop uploading confidential information into random tools simply because the little box looked friendly.

This is not a culture war. It is implementation.

Implementation is where the real work always lives, and that is where the good news sits. If this is change, then we know something about how to manage it.

We know how people adapt

We have already survived the internet, email, smartphones, social media, online payments, digital tax, remote meetings, shared drives, streaming, subscription software, QR codes, apps for everything and the terrifying period when every restaurant decided its menu should be accessed through a sticky laminated square on the table.

We adapted. We complained, obviously, because complaining is an integral part of the British change-management framework. But we adapted, and we will adapt again.

The task is not to decide whether AI is good or bad as though we are reviewing a lasagne. The task is to ask better questions.

What is the work? What is the process? Where is the data? Who owns the decision? Where does the information come from, and where does it go? What should never be uploaded? What requires human review? What could be automated, and what absolutely should not be? What would save time, create risk, improve quality or simply generate faster nonsense?

That is the grown-up conversation. It is less exciting than “the robots are coming,” but significantly less ridiculous.

Beneath the noise is a genuine opportunity—not the fantasy in which everyone becomes ten times more productive overnight and the economy is saved by a chatbot with a friendly tone, but the practical opportunity to redesign work that has become bloated, fragmented and unnecessarily difficult.

We can make information easier to find and complex material easier to understand. We can support people who struggle with blank pages, inaccessible systems or technical language. We can reduce administrative work that drains energy away from human work. We can turn messy conversations into structured next steps, compare options, test ideas, identify gaps and support better decisions.

We can also make expertise more visible and create better bridges between those who understand technology and those who understand the real-world context.

That is where the useful part lives: not within the tool alone, but in the relationship between the tool, the human, the task, the context and the outcome.

That has always been where digital transformation succeeds or fails. It is never only about the system. It is about how that system lands in real life: who uses it, avoids it, benefits from it or is excluded by it; who receives training; who carries the risk; who gets blamed when it goes wrong; and who receives the credit when it goes right.

AI has not changed that. It has simply made it impossible to ignore.

Digital capability is operational infrastructure

For years, many organisations treated digital as a side issue: a website issue, an IT issue, a marketing issue or something for younger employees and “the computer people.” It could supposedly be outsourced, underfunded, patched together and ignored until it broke.

AI makes that position increasingly difficult to maintain.

Digital capability is not a side issue. It is operational infrastructure. It affects governance, risk, productivity, communication, customer experience, staff confidence, commercial value, organisational memory and power.

That is the real shift, and although it is uncomfortable, it is long overdue.

We have expected business owners, employees, boards, volunteers, councils, charities and community organisations to navigate a complex digital environment with patchy training, inconsistent support and the occasional PDF last updated in 2017.

Then AI arrived and everyone acted surprised that people did not know what to do.

Of course they did not. They were not properly supported through the previous five changes either.

That is not a reason to panic. It is a reason to build better support, language, governance, examples, tools, boundaries, questions and leadership. Ideally, it is also a reason to tolerate fewer people selling “AI mastery” after three afternoons and a Canva carousel.

The real skill is learning to think with tools

AI literacy does not need to begin from zero. It can be built upon the digital behaviours people already possess. The task is to explain what has changed.

With search engines, people looked for information. With AI, they can increasingly ask a system to work with information. That is a significant shift, but it is explainable.

Older software often required users to know which button to press. AI tools increasingly require them to understand the outcome they want, the context that matters, the constraints that apply and how to evaluate the response.

That is a different skill, but it is learnable. Frankly, it may be healthier than pretending productivity depends on knowing where Microsoft has hidden the button this year.

The real skill is not simply “using AI,” because that phrase is already too vague to be useful. The real skill is knowing how to think with tools: how to brief, question and challenge them; how to protect data and recognise limitations; and how to use them to support human judgement rather than replace it with beige confidence.

This is not only a skill for technologists. It applies to business owners, managers, assistants, marketers, educators, charities, councils, freelancers, board members, frontline staff and anyone whose work involves information, communication, processes, people or decisions.

Which is, inconveniently, almost everyone.

Messy learning is still learning

The AI conversation is noisy. Much of the content is dreadful, and some posts are so obviously generated that you can practically hear the chatbot clearing its throat before announcing, “In today’s fast-paced digital landscape.”

There are self-appointed AI strategists who could not strategy their way out of a Google Form. The hype, fear and smugness are exhausting.

Yet beneath all that, something useful is happening: people are learning.

They are learning badly, unevenly, publicly and awkwardly, but they are learning. They are testing tools, forming opinions, making mistakes, identifying risks and challenging quality. They are beginning to understand that grammatical output is not automatically good output, and that sounding professional is not the same as saying something meaningful.

They are discovering that automation without judgement is simply a faster route to nonsense.

That is progress, and messy progress still counts.

Nobody becomes digitally capable by reading a policy document and nodding solemnly. Capability develops through trying something, breaking it, becoming confused, asking for help, trying again, recognising the pattern, building confidence and eventually behaving as though you knew it all along.

That is how adoption works. It is not tidy, linear or always flattering, but it works.

We should therefore stop treating this moment as a moral collapse and start recognising it as a capability shift.

The tools, baseline, expectations, risks and opportunities are moving. We need to move as well—not by sprinting blindly into every new platform with a free trial and a suspiciously cheerful onboarding email, nor by banning everything and pretending people will not use it.

We should not allow the loudest person in the room to define the strategy simply because they once asked ChatGPT to write a poem about procurement.

Instead, we should do the boring, useful, grown-up work. We should map processes, understand data, train people, set boundaries, test tools, measure outcomes, review risks and share what works.

We must stop pretending confidence is competence, fear is wisdom or doing nothing is safe.

Doing nothing is also a decision, and often a very expensive one.

The advantage belongs to organisations that understand their work

The organisations that succeed with AI will not necessarily be those with the flashiest tools. They will be the ones that understand their own work.

They will have usable data, clear accountability, relevant training and practical governance. They will also possess enough humility to recognise that adoption is not the same as transformation.

Buying a tool is easy. Changing the way people work is the difficult part, and it always has been.

The good news is that we already know more than we think. We have experienced enough digital change to stop treating every new tool like a meteor. We know that early hype is unreliable, fear travels faster than understanding and poor implementation can ruin good technology.

We know that people need time, support, examples and permission to ask basic questions without being made to feel stupid. We also know that people can adapt at extraordinary speed when the value is clear and the support is real.

That is where Human First Digital sits in this conversation: not within the fantasy that AI will solve everything, nor inside the panic that it will destroy everything, but somewhere far more practical and useful.

This work is about helping people understand the tools, risks, opportunities and decisions in front of them. It is about making digital change legible and giving businesses, teams and communities enough confidence to use emerging technology without surrendering their common sense.

The human element did not disappear because the software became more capable. In fact, it matters more.

The quality of the question matters. The quality of the data and judgement matters. The ability to challenge an output, understand context and ask, “This is impressive, but is it right?” matters.

That is the good news.

AI has not removed the need for human intelligence. It has made the places where human intelligence is missing much more obvious.

Acerbic, perhaps, but useful.

AI did not create the mess

The next phase is not about turning everyone into a technologist. It is about helping people become more literate in the systems they already inhabit.

Digital is no longer a department, a website, a marketing channel or a mysterious concern that belongs with IT. It is part of how work works, and AI is making that impossible to ignore.

So perhaps we can all take a breath.

The robots are not the only issue in the room. The badly labelled spreadsheet, missing access policy, untrained team, mystery CRM, abandoned website, unmanaged inbox, forgotten Google Drive folder and person uploading confidential data into a free tool are all still present.

AI did not create the mess. It has simply made the mess searchable.

In a strange way, that is helpful, because once we can see the mess, we can begin sorting it.

That is the real opportunity—not hype, panic, guru culture or another empty promise that a single tool will repair years of poor processes, weak strategy and underinvestment in people.

The opportunity is to become clearer about work, data, risk, responsibility and value. It is to understand what humans are for in a world where more tasks can be assisted, accelerated or automated.

The answer, thankfully, is not “nothing.”

Humans remain responsible for understanding, deciding, caring, questioning, interpreting, prioritising, challenging, connecting, imagining, persuading, repairing, noticing and leading.

The tools can help, but they cannot be us—which is probably for the best, because judging by some of the AI-generated LinkedIn posts currently doing the rounds, even the machines are tired of pretending everyone is “thrilled to announce” something.

AI is changing things, but things have always changed. We adapted to websites, smartphones, social media, remote work, cloud software and every other layer of digital infrastructure that slowly became normal while everyone was busy complaining.

We will adapt to this as well.

People are not as incapable as the panic suggests. Businesses are not as powerless as the hype implies. AI is not as mysterious as everyone profiting from confusion would like it to appear.

It is tooling: powerful tooling, risky in the wrong hands and useful when applied with thought, context, governance and a clear purpose.

The next time someone asks whether AI is good or bad, perhaps the answer does not need to be a dramatic speech, moral panic or 47-slide presentation.

Perhaps we should ask:

What are you trying to achieve?

What data are you giving the system?

Who is checking the output?

What decision will it support?

What happens if it is wrong?

And does anyone actually understand the process before we automate the chaos?

That is where the grown-up work begins.

Honestly, it is probably where we should have started years ago.