Updated February 2026: the “Risks of AI” section and the closing call to action used fear of inaction as the main argument. We rewrote them - the risks are real, but they call for care, not panic. The section now lists concrete governance questions instead of vague warnings about being outcompeted.
Artificial intelligence (AI) has the potential to change how we work at a pace few technology shifts have matched.
It makes sense to form an early view on which AI tools fit your business, how to use them responsibly, and where to draw the line.
A telling example that OpenAI highlights in its Technical data for the new version of ChatGPT are:
For instance, if multiple banks concurrently rely on GPT-4 to inform their strategic thinking about sources of risks in the macroeconomy, they may inadvertently correlate their decisions and create systemic risks that did not previously exist.”
These new tools can make anyone feel like they have superpowers, but with great power comes great responsibility.
AI can’t (yet) understand how the people in your business work
Below are a couple of things that (generative) AI is not suitable for (yet) and that require actual human intelligence.
- AI has no empathy. Leadership is very much about managing the interaction between people. It requires us to understand the individual drives of people and individuals.
- AI can’t manage creativity, critical thinking and “common sense”. The generative AI that now seeps into all businesses can, to put it simply, only statistically guess its way to what is true, it cannot understand whether something is “true” in the same way as a human being.
- AI is at the moment not a tool to be used to make decisions. AI can radically streamline the way we make decisions, but it needs to be people who make the decisions and have put their human thinking as the basis for the decision.
- AI has very wrong but sounds totally safe.
- Even the newest GPT models don’t know anything about the world after September 2021. We need someone to watch for change requirements of legislators, trade associations and other regulatory entities. You need someone who can understand how to apply it and continuously change approaches as the field develops. For example, by certifying according to ISO 42001 (published December 2023) and tracking how AI Act requirements are applied in your industry.
- The generative AI that has now become very popular, ChatGPT, etc. can only use already done work as a data source, not predicting future opportunities.
Things worth thinking through
Integrating AI into your operations raises questions that are worth taking seriously:
- Legal and ethical compliance: Inadequate AI governance can lead to violations of laws and regulations and damage trust with customers and employees. The EU AI Act (2024/1689) is now in force and sets concrete requirements, particularly for AI systems used in recruitment, credit scoring, and similar areas.
- Data security: AI tools that have access to your data, systems, and communications create an attack surface you need to understand and manage.
- Governance: Employees using AI tools that IT has never approved (“shadow AI”) makes the risk picture hard to oversee.
How can AmpliFlow help you?
- Expert knowledge: Our team consists of both management consultants and technical experts with extensive experience working with companies in various industries. We understand both the technical and operational aspects of using AI and technology responsibly.
- Focus on Ethics and Responsibility: AI should be used responsibly and respect privacy, security and human rights.
- Partner: We see ourselves and our tools as an extension of your team, just as AI can be, and work with you throughout the process, from identifying your needs and challenges to implementing and monitoring solutions.
Get started
AI is already part of your operations, whether you have actively introduced it or not. The question is whether you are governing it or it is governing you.
Book a conversation and we will talk through how to put a reasonable governance structure in place - without making it more complicated than it needs to be.
