Written in the stars: AI and research

Well, the talking heads have been at it again. This last 12 months, we were to invest buckets into artificial intelligence (AI) or be left for dead. Quickly, quickly, more money. No time to waste, they cried. Their sense of urgency, notably naïve, even at the time, their fevered scrawling betraying commercial sensibilities. Though as ever, they were bang on with the blindingly obvious: generative AI is laden with opportunity. It takes my breath away. This industry will both rely on and be thwarted by AI; more than most is. We’ve known that.

As always, the talking heads conveniently forget the sole reason this industry exists – to tell truth apart from hearsay. A bystander might expect this to be natural anathema to an industry which long campaigned for ‘a seat at the table’ because ‘critical thinking, evidence, and methodology matter’. Alas.

Some carefully peppered it into what they had been working on for years, supercharging already intelligent wares, while others supplemented sober human guided automations. Others, in a tellingly swift response, have plied us with months of glitter and pageantry. Those who stare blankly when asked about weighting efficiencies became overnight experts in the appropriate use of large language models. These are the very same who will, quite worryingly, take long hours to fathom the answer to a t-test.

This has elbowed us into a moment when reimagining consumer reaction might be deemed more compelling than actual consumer reaction – where what looks like research is research. No, I’m not making it up – this is actually happening. There are too many readied to believe that querying a language model is (nearly) the same as querying customers. They need so little convincing that five charts and anecdotal case studies will be plenty. Caveats be damned.

So, is this the end? For those betting their livelihoods on this latest generation of snazzy autocomplete, yes.  Yet for those who appreciate this is but a temporary hold, prepared to choose their moment and stay true to purpose, a universe awaits still.

Previously, I shared the four hallmarks of profitable innovation within insights. They remain: consolidation, responsibility, iteration and creative assembly. In swathes of innovations, none are manifest. A link into OpenAI is not this.

To those agency leads who did not go with the flow and are still sweating it out, white knuckled at the wheel, my message to you is that your convictions were sound, for history is indeed repeating itself. The smartest commercial moves remain, already well-rehearsed by those who didn’t embarrass themselves through the last wave of AI.

They never pretended to invent something that was already a dusty toy on big tech’s shelf. They weren’t the first at the party – no one invited them, you didn’t even know who they were, but when they came knocking, they knew why. They moved like lightning at a fraction of the cost.

Today, they sit silently, once again transfixed on dark skies, waiting for the GitHub stars to align. They will leave the open-source community time to do so, working with them, ensuring smart ideas become quicker and slicker.

Then some stars will twinkle, swell and swallow others by force of gravity, and you’ll surely miss it if you’re not watching.

If you are, you know this dance has begun. Now certain that your competition has, in short months, traded their war chest for fools’ gold – a messy repository that we will replace with three sweet lines of Python. Worse still, they have hitched their wagon to the most banal and misguided possibilities of a fearsome advancement. It cannot be any other way; to say the least, the fundamental floorboards of these models remain in flux.

This means every week resisted saves long months, if not years of development time. So, you must suffer the pageantry. Smile at the party. Hold your nerve until the very second AI tells more insightful real stories to replace believable fake ones. Until you see the flash, steel yourself through our darkest hour. Your time to shine draws ever near.

Ryan Howard is a freelance marketing science consultant