Why urgency without strategy can undermine trust in the technology.
Introduction
If one of the biggest barriers to traditional technology adoption was analysis paralysis and an inability to decide, the rise of generative AI and autonomous models has introduced the opposite extreme: a compulsion to decide without criteria. Under pressure from the narrative of unstoppable acceleration, many organisations are skipping strategic alignment and creating a wave of superficial implementations that lead to internal frustration and declining external trust.
Incentives and technology
The perception that AI is advancing at an unprecedented pace is not a marketing myth. It is a structural dynamic driven by incentives that are too closely aligned to be easily resisted:
- Geopolitical level: The race for technological supremacy between superpowers such as the United States and China turns AI development into a matter of sovereignty and national security, removing any incentive to slow down.
- Technology provider economics: The market for AI infrastructure and models operates on a runaway-leader logic in which the winner captures a disproportionate share of value, pushing technology providers to invest and release capabilities at exceptional speed.
Against this backdrop, Jacob Coxon, a former pre-training researcher at OpenAI and Anthropic, spoke publicly about the pace of AI progress, the accelerating capability of models and the risk of superintelligence emerging through a singularity or intelligence explosion.
In an interview with CNN, Coxon described how models had moved very quickly from assistants towards autonomous systems capable of outperforming people in complex tasks:
“About two years ago, these AI systems served primarily to help people. They could give suggestions. Now they are close to replacing them… It’s quite plausible that, within a year, we will no longer need people to do research in many areas.”
He added that humans could plausibly become unnecessary for research in many areas within a year.
Coxon also explained the transition to recursive self-improvement, the mechanism at the centre of the singularity concept:
“Pode dar-se a um modelo de IA o problema da investigação em IA. E, então, obtém-se o que se chama uma explosão de inteligência.”
From fear of falling behind to adoption fatigue
When this rush to adopt enters the corporate environment without mediation, it triggers an impulsive response: adopting tools simply out of fear of losing the race, which we call FOMO – (Fear Of Missing Out).
This haste can produce two immediate and damaging effects:
- Loss of credibility with employees: When teams are exposed to unstable tools without training, integration into real workflows or a clear connection to genuine pain points, they develop scepticism and “innovation fatigue”. AI starts to look like an administrative burden or a leadership public-relations exercise rather than a productivity amplifier.
- Erosion of customer trust: Customer-facing solutions launched before they are mature — such as ineffective virtual assistants or failed customer-support automation — weaken the value proposition and expose the organisation to operational and reputational risk.
Building balance: decision models for disruption
Overcoming this dilemma requires recognising that rapid technological progress does not justify rushed corporate decisions. Leadership does not need to follow every market iteration. It needs to build a balanced decision model anchored in five dimensions:
- Strategic alignment: Evaluate the contribution of each initiative to business objectives in order to ensure the selection of those with real value and measurable impact (avoiding generic AI applications).
- Human factor: For each initiative, assess cultural maturity, openness to change, the digital capabilities of the employees involved and the presence of a leader who can catalyse transformation — particularly important in the early stages of adoption.
- Predominant AI technology: Map the technology architecture of each initiative to promote economies of scope, reuse and internal specialisation instead of accumulating disconnected tools.
- Data characteristics: Diagnose the availability, history, accessibility and quality of the data required to operationalise AI models.
- Risk assessment: Measure the sensitivity of the information being processed, tolerance for unexpected outcomes, and the potential reputational and operational impact.
From reaction to considered leadership
If hesitation paralysed companies in the past, haste can destabilise them in today’s AI ecosystem. True digital maturity is not measured by how quickly an organisation adopts the latest development, but by its ability to make clear, independent investment decisions focused on sustainable value creation.
