True, ATCOM has been at the very forefront of AI development and implementation for a while now – actually, since it was, rather mundanely, called ‘machine learning’. Indeed, we have been excited to evolve in lockstep with AI, integrating it increasingly in the services we offer to clients – and to observe it delivering concrete results, with the optimization iterative loop amplifying them further. We have seen both simpler, internal benefits and value creation materialize: in the former, cost savings, greater speed and improved scale capacity (say, through the smart repackaging of content) is being delivered; in the latter, creative work is enhanced while, at the same time, customers are served improved experiences through new generation chatbots or superior personalization.

And yet we believe that the current understanding of (especially generative) AI is in danger of entering a perilous hype phase, with the usual disappointing (or, on occasion, disastrous) results that such hyperbole entails – indeed, when it comes to introducing, integrating and implementing AI into clients’ offerings, we think that a proverb from more than a couple of millennia ago, is quite pertinent: attributed to Chilon of Sparta, one of Antiquity’s Seven Wise Men, σπεύδε βραδέως (to ‘make haste slowly’ or ‘hurry slowly’) embodies the imperative to be distinctly cautious when moving fast, to avoid errors that come from excessive speed. It is, in a way, the opposite of Zuckerberg’s “move fast and break things” – a brief reflection on the major ills (alongside its numerous benefits, needless to say) that the current social media ecosystem is plagued by certainly proves the point.

And, truth be told, along with an abundance of opportunities, generative AI’s path is beset with hazardous risks – take for instance data security and privacy, as such models often require vast amounts of data for training and optimization. If organizations hastily adopt these systems without robust safeguards, they risk exposing sensitive customer or proprietary information to breaches or misuse - such incidents can lead to reputational damage, regulatory fines and loss of customer trust. Or consider ethical and legal challenges: generative AI may create content that is misleading, biased or infringing on intellectual property rights. Companies deploying these tools without adequate controls may face lawsuits, regulatory scrutiny or public backlash.

The issue of trust cannot be overstated – simply put, it underlies all endeavors and constitutes the ultimate success factor for any undertaking. It is multidimensional too: we at ATCOM use the term ‘Dual Trust’ as there are two distinct constituencies to be addressed, one internal and the other external – the former is the company’s own talent that need to be reassured that AI will enhance rather than replace their roles; the latter is the end user that must have full confidence that his or her data are both secure and will be deployed in a transparent manner, with full explicit consent. Fail here and plans come to naught – or worse. Succeed and then (and only then) possibilities are endless indeed.