The Algorithm Has No Taste

In modern life, time has become our most precious commodity - and an absurd amount of it is steadily consumed by the admin of other people's milestones.

Consider the sheer volume of occasions that demand your attention across a single year. There are the birthdays for children, partners, parents, siblings, nieces, nephews, close friends and god children. There is Christmas, which seems to be an endless doom scroll from late November to mid December. There is the relentless stream of your children's classmates' birthday parties. And peppered between it all are teacher gifts, office Secret Santas, last-minute thank you’s, get well soon’s and congratulations token gifts to say that you’re thinking of them.

Each one requires you to stop, think, search, compare, decide, and act.

• Individually, these tasks might seem manageable - a quick bit of life admin. Collectively, they represent a recurring drain on your time and your mental energy. Research confirms that the cumulative weight of repeated decision-making leads to "decision fatigue," a state in which our cognitive performance deteriorates and the quality of our choices takes a nosedive. Gift-buying is a near-perfect example: the more occasions you face, the more likely you are to default to something safe, generic, and instantly forgettable - the Jo Malone candle or the Ottelenghi cook book, simply because your capacity to think creatively or carefully has already been spent elsewhere.
• At the same time, we are living through a fundamental shift in how we seek answers. The traditional Google search experience is being rapidly dismantled, replaced by AI-powered summaries that now squat at the top of our screens. These tools are changing behaviour at speed. The appeal is entirely obvious: AI promises a frictionless "quick hit", sparing us the tedium and time black hole of doom scrolling through multiple websites.

But if time is a luxury, it is worth asking what exactly we are sacrificing when we outsource our creative thinking to an algorithm?

From Blue Links to Bland Summaries

To understand the illusion of the "quick hit", you have to look under the bonnet at how search has changed, and what that means for an everyday task like buying a gift. The mechanics dictate the quality of what ends up in our shopping baskets.

The Old World: Pre-AI Google

• Historically, Google was just a very efficient filing clerk. It crawled the web and ranked pages based on relevance and authority. If you searched for "best birthday gift for a 30-year-old amateur chef", it handed you a list of blue links pointing to articles and shops. The burden of synthesis remained entirely on you. You had to open five tabs, cross-reference the reviews, and make the call. The glaring weakness, of course, was that the top results were heavily rigged by Search Engine Optimisation (SEO); the brands with the biggest marketing budgets dominated page one, effectively burying anything small, independent, or unique.

The New World: Google AI Overviews

• Today, Google has bolted on its Gemini language model to create "AI Overviews". It now acts as a sort of frantic research librarian, taking your query, running multiple concurrent searches, and synthesising the top pages into a single, readable summary. It saves you clicking, but it inherits all the old biases. The AI is simply summarising the most popular, highly-ranked pages. Search for that chef's gift, and it will confidently recommend a mass-market knife brand simply because it appears on the most affiliate marketing lists. It offers the illusion of a definitive answer, but it is really just an echo chamber of the internet's most heavily marketed products.

The Wildcard: ChatGPT, Claude, Perplexity and AI Assistants

• AI tools and assistants represent a shift from retrieving information to completing a task. When you ask it for gift ideas, it appears to act as a creative consultant, generating a response by predicting which words should logically follow your prompt based on its vast training data. The problem is extreme inconsistency and a glaring lack of true creativity personalisation. Recent research shows that when asked to recommend products, AI tools pull from what is essentially a "statistical lottery". You might get a nice cutting board, or a popular cooking class. But even with a highly specific prompt, ChatGPT lacks the human intuition to know that your friend actually prefers vintage, hand-forged Japanese steel. It cannot replicate taste.

The Time Trade-Off: Speed Gained, Quality Lost

The appeal of AI search is straightforward: it saves time. But it is worth understanding precisely how much time is being saved, and what is being surrendered in exchange.

With traditional Google search, finding the right gift was a research exercise. A shopper searching for a birthday present for a close friend would typically open multiple tabs, scan several "best gift" listicles, cross-reference reviews on Amazon, check a brand's website, and perhaps consult a forum or two. Research shows that the majority of online shoppers - 61% - spend between 10 minutes and an hour researching before making a purchase, with a significant proportion spending considerably longer for considered or personal gifts. For a meaningful present, that process could easily stretch to 30 to 45 minutes of active searching, reading, and comparing.

With Google's AI Overview, that same search now returns a synthesised summary in seconds. The AI reads the top-ranking pages on your behalf and presents a tidy list at the top of the screen. The time investment drops from 30 minutes to perhaps 30 seconds. With ChatGPT, the process feels even more conversational and immediate - type a prompt, receive a list, done.

On the surface, this looks like an extraordinary gain. In reality, it is a trade of depth for speed. The time saved is real, but what fills that time saving is a result built from the most popular, most marketed, most algorithmically prominent options on the internet - not the most thoughtful, most personal, or most unique.

• A 2025 survey of 2,000 consumers found that 60% of shoppers already feel that AI makes the gift-buying process feel "more mechanical". That instinct is correct. It’s like the feeling you get after panic buying a gift on Amazon - often a wave of guilt and shame. The algorithm is optimising for efficiency, not for the emotional resonance of a truly considered gift.

The result is entirely predictable: a generic list. Search for a birthday gift for a 40-year-old who loves cooking, and every platform - old Google, new Google, and ChatGPT alike - will converge on the same shortlist. A Le Creuset casserole dish. An Ottolenghi cookbook. A nice set of knives. These are fine gifts. They are also the gifts that thousands of other people will give this year, because they are the gifts that dominate affiliate marketing lists, SEO-optimised gift guides, and AI training data in equal measure.

The time saved has come at the cost of originality. This is the vanilla-taste trap that many are falling into.

The Limits of the Algorithm

When you try to use AI for tasks that require actual human nuance, the limitations become glaring. Research from MIT points out that Generative AI is only ever as good as the prompt you feed it. But even with the perfect prompt, an algorithm hits a wall. It operates on predictions and probabilities based on past data; it does not possess emotional intelligence, intuition, or taste.

This is particularly obvious when it comes to giving gifts. A study in Psychology & Marketing found that consumers actively resist using AI recommendation tools when buying for close friends. Gift-giving is a deeply social act - it is meant to signal that you know someone, that you understand their preferences. An algorithm might suggest a popular brand based on keywords, but it cannot replicate the empathy required to uncover something genuinely unique.

The Case for the Human in the Loop

At BlckBx, we know exactly where AI fails, which is why we know exactly where to deploy it. We use it for the heavy lifting: parsing massive amounts of information, identifying trends, and aggregating initial research. But we do not let it make the final decision, and we certainly do not let it dictate taste. Our approach is to harness the raw processing power of AI, marry it with a proprietary data graph built around a client's personal context, and then put a human expert in charge of the outcome.

This hybrid approach is backed up by the data. The MIT Center for Collective Intelligence found that while AI is brilliant at repetitive, data-driven tasks, humans remain vastly superior at anything requiring contextual understanding and emotional intelligence. Put the two together, and the outcomes far surpass either one working alone.

Behind the BlckBx platform sits a specialist gift-buying team whose knowledge goes well beyond anything an algorithm can access. These are people who follow trends closely, who understand what elevates one brand above another and why, and who know the small, independent labels that never make it onto a gift guide. They know where to find the sustainable makers doing important and interesting work, the vintage dealers with one-of-a-kind pieces, and the second-hand sources that can be trusted. This is knowledge that can’t be prompted out of a chatbot. It is built over time, through human experience and intuition, curiosity, and taste that is the difference between an excellent gift and an instantly forgettable one.

At BlckBx, human expertise and verification is the safety net. Every single task output executed for a client - whether it is a unique gift or a complex holiday itinerary - is rigorously reviewed and verified by a human assistant before it sees the light of day.

By keeping a human in the loop, we ensure the final recommendation is a curated, thoughtful choice, rather than just a statistical probability.