Why Gordon Ramsay might just be a better chatbot or robot than Jamie Oliver… for now.
Friday 30th June
In the run-up to London Food Tech Week this October, Chris will be sharing his perspectives on each of the week’s key Day Themes. This month he starts with our day focussed on the The Tech Food Reality; where he explores the applications of artificial intelligence and chatbots used in other sectors for inspiration on how they can be used in the kitchen.
Picture this scenario… Friends are coming over for an impromptu early dinner this evening and you need to sort dinner fast. Coming to the rescue is your Internet of Things (IoT) connected Siemens smart fridge that shows you a photo on your phone of exactly what you’ve got left in your fridge.
You ask IBM Chef Watson via voice control to recommend a vegetarian starter that will make use of the extra asparagus, sage and fennel still sitting idly in your fridge screaming out to be used!
You choose the benign sounding, but slightly out there “Asparagus Broth” recommended by Chef Watson based on his novel AI-driven interpretation of a once classically composed recipe that will now use “ale, asparagus, fennel, olive oil, horseradish root, gorgonzola, sage, cilantro, marjoram, star anise, and turmeric” to make an inspired broth you are sure no one else has ever tasted before.
Your chatbot knows what’s missing from your fridge and goes about adding the rest of the ingredients to your existing shopping list on Ocado. You agree to a delivery slot of 2:30pm this afternoon knowing with certainty it will be delivered by hook or by drone.
The main course is Gordon Ramsay’s Beef Wellington recipe and once the Ocado delivery has arrived you will have Ramsay’s chefbot in your kitchen directing you via your AR Microsoft Holo-Lens glasses on how to cook it to perfection – yes, Gordon’s autocratic style may leave something to be desired but there’s nothing like being shouted at by the man himself! You get to feel you’re right there in the heat of it – Yes, Chef!
Dessert is a relative doddle – a 3D food printer will work out the nutritional content of the first two courses and calculate the remainder of our daily personal dietary requirements suggesting two beautifully constructed nautilus shaped desserts nutritionally balanced and supplemented with the latest in custom-blended vitamins and herbal superfoods designed to boost vitality and mood.
As dinner closes, you pull out your chatbot-operated Bonaverde coffee machine, which lightly roasts some Panamanian beans you have in your kitchen, grinds them and pours out a perfect pot of coffee within 10 min – the perfect way to end the dinner.
For those of us in the Food Tech space – this picture of the future is exciting and there has been sufficient progress and enough examples to suggest that this future is an inevitability from a technical perspective. However, as with most tech related things, the question is how far away is it actually from being a reality and also how will we, as humans with our quirks and ingrained patterns of behaviour, embrace these changes?
One of the key technologies that allows the above scenario using Internet of Things (IoT) devices to function seamlessly is the use of AI-based chatbots – both of the voice variety and the text-based messenger type.
With the spread of Alexa, Google Home and even the long-suffering Siri – many of us have actually experienced the frustration of asking relatively simple questions that these AI chatbots just don’t understand but any 8 year old would. Try asking Alexa “Should I add more salt?” and hear her squirm, “Sorry I don’t know that one”!
While the technology is improving on a daily basis and no doubt will eventually get there, at this point in time, there are only a few use cases that a simple chatbot can handle and those primarily revolve around efficiency questions and repeatable tasks – more a singular response to specific questions or search queries with predetermined outcomes or answers, for example asking Alexa for a weather report.
This isn’t real engagement and not really a chatbot in the sense of the word – ‘chatbot’ implies an element of continuing two-way dialogue with strong Natural Language Processing (NLP) abilities and here a view of our current limitations to the tech is interesting.
Given these bumps in the road towards tech enabled utopia, are there any examples today of where AI-driven chatbot technology works so well that it is indistinguishable from a human interaction? And if so, how might we apply it to the food sector?
I think the answer comes from understanding the social dynamics of the only examples I have come across to date whereby a chatbot is able to act effectively as a human surrogate. Interestingly they have come out of conversations from outside the food and hospitality sector.
The first use case is an EdTech start-up called Eggbun, a chatbot tutor specifically for Asian languages whereby an English-speaking student is taught by an AI-driven chatbot to learn Korean, Japanese or Chinese. It has been tested to work incredibly effectively compared to any other type of language programme and has been very successful with significantly higher user retention and continued usage rates after 6 months.
It’s a service that I could even see myself using, which has driven me to think about what it is that makes it work so well. In short, it’s because the company has cleverly flipped the natural order around of how we humans generally operate with machines.
We humans are used to being the High status partner when it comes to our relationship with robots/machines who are built to serve us (Isaac Asimov’s 3 Laws of Robotics are probably subconsciously guiding us here) – but with that comes expectations and ‘unreasonable’ demands on the robots that they must respond to our whims. However, in the case of Eggbun this relationship is flipped around and the human now occupies a position of Low status, as a student with limited vocabulary and ability to express themselves, and the chatbot occupies a High status position We now know our place and all of a sudden, the technology works beautifully and we barely mind that it isn’t human.
The second use case I came across reaffirmed this High status-Low status relationship and came about through a conversation with a Chinese tech investor. I was asking about how advanced the tech scene in China was compared to Europe and he shared an example where the chatbot was advanced enough that the customers using it all believed it to be a human and not for a moment realising it was an artificial entity.
This use case was a chatbot deployed in the Fintech sector for a money lender/loan provider whose chatbot would call customers who were late on their payments to ensure they complied and paid the arrears and instalments on time in the future. The voice chatbot was even programmed with a North Eastern Chinese accent known for its slightly menacing tone that would encourage the customer to pay up promptly.
In this case, the reverse High-Low status between human and machine is in play again and is, I believe, a key reason for its effectiveness. By giving the chatbot a firm and menacing demeanor, the Low status entity, i.e. the human, no longer questions authority and asks challenging questions. This is important because there are only a very limited number of responses that can reasonably be covered by the AI so reducing the chance of difficult questions being asked is a large factor in its success.
These examples provide really interesting use cases for chatbots and open up a new understanding of how we can engage with and relate to AIs. What this means for the future of chatbots in Food Tech I don’t yet know but I invite entrepreneurs and innovators to explore this relationship model and see how it could be applied effectively.
Referring back to the hypothetical dinner party I described above, this High-Low status dynamic suggests that now could be the time for a Gordon Ramsay recipe chatbot – one where we could conceivably believe we are interacting with a hot headed and foul-mouthed version of Gordon himself on the set of Hell’s Kitchen. For the Jamie Oliver fans out there, you may have to wait a little longer to find a friendly, cheeky chappie Equal status chatbot to convince us.
London Food Tech Week has five different day themes, which hospitality tech expert and YFood contributor Chris Fung will be exploring over the next few months. You can find out more about each Day Theme here and get your hands on a Week Pass, which will be the only passport you need to access the speakers, content and events that interest you the most during the Week.
Subscribe to our mailing list for all the latest news from the Food Tech space. Never miss out on updates on our monthly Food Tech Wednesdays meetups and our next Food Tech Week.