Digital assistants can support us with our daily tasks and reduce our workload in various ways. This makes them interesting targets for businesses that want to add us to their customer base. Will we soon be controlled by AI?
“Good morning! You did not sleep well, your heart rate was higher than usual. You will become tired at two o’clock this afternoon, and your meeting with the project team is scheduled for half past two. I recommend that you take an iron supplement at noon. I will now start playing the ‘Groove’ playlist to wake you up. I have already boiled the kettle. Have a cup of black tea and do not let it steep for longer than two minutes. This will have a stimulatory effect. Today the weather is cloudy, but no rain is forecast. The temperature will not exceed 18 degrees. The refrigerator is signaling that you are running low on milk – shall I order two cartons?”
The voice coming from my smart speaker is the voice of my metabot. It is my very own personal assistant, which can access countless data that I have entrusted to the internet. It is Siri, Alexa and Hey Google in one. It manages my schedule as well as my fitness, keeps me entertained, communicates with my kitchen appliances and media gadgets, and takes care of tedious everyday tasks. And in doing so, it learns more and more about me. It might already know me better than I know myself, which is why, as a customer in the global marketplace, I’m no longer as important as I used to be.
Instead, many businesses are far more interested in my metabot. That is who they have to get on their side, that – or more precisely, its AI – is who they have to pitch their products to. And they have to make sure that their brand is as high up on its list as possible. “Businesses are anticipating this development and have already started to adapt their marketing activities to metabots,” says Anne Scherer, assistant professor of quantitative marketing at the Department of Business Administration.
If (question), then (answer)
Scherer believes that metabots will be a reality in the United States in about 15 years and thus shift the paradigm in the field of marketing. It will likely take a bit longer until this happens in Switzerland, but with chatbots such as Siri and Google Assistant, we’re definitely headed in this direction. However, the decision-making and language skills of many chatbots are currently based on relatively simple dialogue trees – if (question), then (answer). The metabots of the future will also be able to understand context, regardless of platform-specific areas such as music or shopping. But here’s the catch – metabots accumulate personal data, but don’t always do so in the interest of the customer. This means that there’s a risk that they will give businesses more market power than the customer is comfortable with. For example, it has been empirically proven that Amazon’s chatbot Alexa prefers to recommend products from the company’s own AmazonBasics range. A German car company is working on a similar idea: If there’s a problem with the vehicle, they want the on-board computer of their pricey cars to recommend a garage of their choosing rather than the one that’s closest. So can AI be trusted? “Strictly speaking, we should read the general terms and conditions of every single app before we use it,” says Anne Scherer. However, the average internet user would need 76 working days per year to comb through the GTCs of all the services they use. “When it comes to the small print, businesses have taken the easy way out,” criticizes Scherer. “They shouldn’t pass on responsibility to the customer, but be more proactive in taking it on, for example by increasing transparency and giving users more say.” Since companies aren’t doing this, the use of AI requires rules, and this is the area in which Scherer specializes. For an interdisciplinary TA-SWISS study exploring the risks and opportunities of AI technology, Anne Scherer wrote the chapter about AI and consumerism. AI – and thus also the digital marketplace – can be fairly obscure to the uninitiated, like a black box. Most internet users aren’t aware of what they’re revealing by sharing their data. What’s so important about my Facebook post? Who cares that I’ve liked the new Star Wars page? But AI remembers all of our activities and learns in the background. The algorithms are programmed to compare data and draw conclusions from them. The more data they have, the more intelligent the machine. And anyone who has seen Stanley Kubrick’s 2001: A Space Odyssey doesn’t need to be reminded that smart machines can be dangerous. Nevertheless, Anne Scherer believes it would be wrong to regulate and possibly slow down the technology itself. “An actual AI law would be wrong. Regulation must begin with consumer and data protection, especially in the gray area of insights derived from algorithms, for example when it comes to people’s emotional state and sexuality.
Is AI making us dumber?
Dynamic personality profiles have already begun to replace conventional static ones. A few hundred likes on Facebook or a simple profile picture are all the machine needs to generate and exploit a user’s private traits and attributes with remarkable accuracy. Over time, self-learning AI thus enables businesses to estimate their customers’ current state of mind and focus on the much-vaunted customer experience rather the customer’s actual needs. Where will it end? Will metabots and AI really make our lives easier? Anne Scherer nods. Businesses can speed up the product development and sales process and get feedback on the quality of their products through the users’ data. What’s more, this feedback also eliminates the need for expensive market research activities. On the customers’ side, AI will help them navigate the vast digital marketplace and make personalized decisions. “AI can inspire, activate and help me. It can direct and focus my attention. However,” says Scherer, “AI isn’t as intelligent as many people believe, because it is trained to recognize patterns. Intuition, abstraction and imagination are and will remain human strengths – for now, at least.” The question to ask is whether and how human strengths will change by having AI in our lives and metabots telling us how long we should let our black tea steep. The concept of outsourcing one’s memory is already a thing. Since all kinds of information – from the local bus timetable to facts about world history – is available online at all times, it is no longer saved in our brains. Does this mean that people are getting dumber? Not necessarily, since this frees us up for other, possibly more creative tasks. Actually, this isn’t the first time that our memory has been “outsourced”: Back in the 15th century, the printed word replaced oral tradition. And this certainly didn’t hurt the world.