It's been over a year since ChatGPT was released in November 2022, kicking off a wave of hype around AI technologies to rival the early internet. But today, for all the hype, investment, and discussion around AI, the world hasn't quite changed yet. Many of the demos we've seen of the capabilities of AI for consumers have been scaled back significantly, and many early entries into the new market of AI products have struggled to gain traction.

Some of these products are clearly impractical, borderline silly attempts to rush out half-baked products. These attempts aren't stopping anytime soon, and consumers will continue to be bombarded with AI-everything. A lot of these features are valid, legitimate improvements, but others may bring nothing but lofty expectations and a vacuous sense of failure. With all the extra pushing, it won't be surprising if people get burned out on the whole idea.

AI is a classic hype-cycle

The future will involve AI, but it might take some time

The Gartner Hype Cycle is a stalwart feature of discussions about the future of tech. It's an interpretation of how excitement around new technology quickly peaks with a 'trigger', a proof-of-concept that demonstrates the limitless potential of a new technology, gathering intense excitement only to fade as implementations struggle to make a significant impact. Over time, the viability of a technology improves quietly in the background, and it slowly gains more traction and is widely implemented, often without the fanfare it initially received. While this interpretation is open to criticism, and certainly has its flaws, it's easy to follow some examples through. Augmented reality is a good example of a technology with a huge hype cycle (remember Google Glass or SnapChat spectacles?) but that is now starting to become more relevant with the Apple Vision Pro and all kinds of AR glasses. Quantum Computing is another example that perhaps hasn't yet had its trigger point moment, launching it into the public consciousness and triggering a run of adoption and investment.

AI certainly feels like it's now had its 'trigger point' and is at the peak of expectations. The technology is clearly incredibly powerful, but the examples of its use we've seen so far have been limited, or still have big caveats or problems to overcome. If history tells us anything, it's that excitement around AI will fade, and maybe we should be expecting a slower burn to true consumer adoption. This kind of consumer burnout has happened before, and applies to almost everything; however exciting something is, it can struggle to stay in the forefront of the consumer mind for long.

Products will remain limited for consumers

CES 2024 was the year of AI, but was it?

While there are some interesting uses of AI coming out by companies, most of the offerings for practical consumer products are weak at best. For most consumers, we don't necessarily need nor want endless AI chatbots, especially if they fail to add any tangible value. This is where AI may risk a burnout with consumer patience, as excitement around the technology itself struggles to translate into genuinely useful and cost-effective products for consumers.

Take CES this year, which saw a wave of new AI products announced; most of which were at best thin attempts to ride the hype wave. We saw AI-powered lighting and mood monitoring, as well as voice-controlled AI-powered vacuum cleaners and even an AI-powered butler-bot from LG. While novel and potentially useful, these products arguably represent a problem of unfulfilled expectation for consumer AI.

The concern for companies should be making tasteful use of AI, using its capabilities in cost-effective ways to improve existing products. Overloading every item with voice assistants is likely to become not only frustrating and exhausting for consumers, but a bad long-term decision for both businesses and consumers (as many of these devices will need ongoing enterprise subscriptions to underlying third party models, or compute to run.) Injecting AI novelties may also dampen expectations of its potential before it's realized. We saw this with 'smart' devices — a wave of internet-connected fridges, ovens, washing machines and other household appliances with often poor functionality, little usefulness, reliability or a host of other problems that has harmed the entire image of smart devices in the home, potentially even leading consumers to avoid such devices in the future.

But have businesses successfully adopted AI?

Small, quiet improvements have been successful

The companies making the best use of AI are doing it in small ways, making efficiencies or improvements to their services without overhauling the entire product into an AI blackbox. One of the big stories of AI's use so far is in replacing online customer-service agents with AI chatbots capable of handling customer service inquiries and even resolving problems on behalf of customers. Companies like Octopus-Energy, a disruptive UK energy supplier, have been utilizing ChatGPT to replace the workload of over 200 employees. Klarna, an early banking partner for ChatGPT enterprise, has also recently said they're expecting to see $40m of cost reductions through the integration of AI-powered customer service representatives, which are outperforming their human counterparts on a number of metrics. Coca-Cola is also making use of Dall-E for marketing and customer experience.

It's not all a success story though; Air Canada was recently held liable after one of their chatbots gave a customer incorrect information about bereavement fares.

We're already using AI every day

It's also easy to forget that despite this new wave of hype generated by LLMs, we're already using AI every day, whether you realize it or not. AI is already in use for everything from weather predictions to approving mortgages. Your phone is already likely to use AI features to touch up photos or suggest search results. LLMs have made our interactions with AI easier and more tactile for everyday users, but there's plenty of AI in use all around us already. It might not be as flashy as LLMs, but it's equally essential in how it helps companies develop data-driven decision-making.

Consumer AI will be about iteration, not revolution

Most of the capabilities of AI in consumer devices will represent enhancements, not revolution, at least for now. Not every device in your life needs a voice assistant. For now, the face of consumer AI will be in improving data, helping to track sleep patterns, offering lifestyle insights, or setting room profiles for speakers. It will be in improved tools for education or searching the internet, enhancing our interactions with technology without revolutionizing them. But we may have to wait to see this potential truly realized, and in the interim, it may be businesses who are able to take true advantage of the potential of AI.