It’s a bit ironic, isn’t it.
In the context of ‘personalization’ as a marketing trend, technology – specifically Artificial Intelligence – is used to make digital experiences more personal, more people-focused.
Instead of arriving on a generic landing page, with tech-driven personalization, you arrive on a page that speaks to you with the recognition of a human salesperson. “Hi Robin, ready to buy those round trip tickets to Boston?” a travel site’s page headline might say, using my browsing history, past forms filled, and other data to guess at my purpose. Pop-ins could follow me as I explore their site, reminding me about sales to destinations I’ve previously shown interest in, like an astute travel agent. The cherry on top: an SMS that uses my name to confirm my payment, maybe even sent by the company chatbot, Steve, wishing me bon voyage.
Essentially, ‘personalization’ is an attempt to use technology to anthropomorphize our digital interactions.
Is there any kind of inherent contradiction there? What’s the current state of the human vs machine relationship in the Algorithmic Age? How good of a job are these algorithms really doing? And finally, is ‘human-centered’, AI-powered personalization really possible?
How AI is piloting personalization
“Personalization’s image of us is like looking at yourself in the funfair’s house of mirrors. Personalization caricaturizes us and creates a striking gap between our real interests and their digital reflection.”
He was referring to his so-called ‘paradox of personalization,’ summarized below – that AI-powered personalization aims to respond to an individual’s personal predilections, while at the same time striving to shape and direct them. This means that consumers are still sometimes bombarded with irrelevant ads, personal AI assistants don’t always get it right, and we often get lost in newsfeed ‘echo chambers.’ Though progress has surely been made, I wouldn’t totally disagree with that assessment in 2018.
“Additionally, there lies a more general paradox at the very heart of personalization.
Personalization promises to modify your digital experience based on your personal interests and preferences. Simultaneously, personalization is used to shape you, to influence you and guide your everyday choices and actions. Inaccessible and incomprehensible algorithms make autonomous decisions on your behalf. They reduce the amount of visible choices, thus restricting your personal agency.”
Part of the problem is that algorithms face inherent limitations – from the constraints of the datasets used to train them and finite computing power to the conflicting interests of those designing the algorithms versus those consuming them. It’s clear that, when it comes to AI-based personalization, there’s always room for improvement.
Ambivalent feelings about AI
Jarno’s assessment embodies society’s ambivalence regarding the increasing prominence of AI in every aspect of our lives, whether it be in marketing personalization techniques or in government, healthcare, media or anywhere else.
To wit, in 2017, Pew Research Center asked 1,302 technology experts, scholars, corporate practitioners and government leaders: Will the net overall effect of algorithms be positive for individuals and society or negative for individuals and society? Respondents seemed torn: 38% predicted that the positive impacts of algorithms will outweigh negatives, while 37% said negatives will outweigh positives and 25% said the overall impact of algorithms will be about 50-50. The themes teased out of their answers shows a recognition of algorithms’ potential blessings as well as pitfalls for the human individual:
When it comes to ‘deep’ AI, à la Terminator or Blade Runner – that is, true super intelligence that mimics and surpases humans’ own – people’s opposing feelings intensify.
Stephen Hawking famously told the BBC, “The development of full artificial intelligence could spell the end of the human race.” The Future of Life Institute, an organization dedicated to catalyzing humanity-first research efforts, and of which Stephen Hawking and Elon Musk are members, felt the need to establish 23 ‘Asilomar AI Principles’ to help guide the future of AI. These include ethical considerations like principle 11, Human Values: AI systems should be designed and operated so as to be compatible with ideals of human dignity, rights, freedoms, and cultural diversity, and principle sixteen, Human Control: Humans should choose how and whether to delegate decisions to AI systems, to accomplish human-chosen objectives.
Other thinkers are much less pessimistic, ready to announce the significant benefits super smart robots will bring to society. Ray Kurzweil, chief engineer for Google, futurist and well-known AI authority, consistently maintains that the latter will help humanity progress. A year ago, he told the Council on Foreign Relations that the Singularity – a term coined to denote the point at which AI will surpass collective human intelligence – will enable an augmentation of our abilities through a sort of ‘interbreeding’ of man and machine. Indeed, we might not need to wait that long for hybrid human/AI creations, as organizations like the Defense Advanced Research Projects Agency (DARPA) are only too ready to announce.
Different and complementary?
Ask any Tom, Dick or Harry off the street, and they’ll probably tell you that AI is good at some things, but couldn’t possibly be better – at least any time soon – at all things, especially those that require a lot of interpersonal interaction and emotional intelligence. Looking again at work from the Pew Research Center, though Americans think robots might soon displace fast food, insurance claim and software engineer workers, they’re more skeptical that an algorithm can do the job of a teacher or a nurse. With a nod to quintessential human narcissism, most also found it unlikely that their own profession would be automated away.
Though it might seem obvious that algorithms can’t compete with humans on subjects relating to emotions, the arts or humanities, a few cases in point prove that the question isn’t black or white. The Turing Tests in the Creative Arts, held in 2016 at Dartmouth College, aimed to see whether ‘machines could create artwork in specific contexts indistinguishable from human-generated work.’ They found out that they could – Christopher Raphael, a professor of informatics and adjunct professor of cognitive science, designed a program that wrote a song which judges found to be more or less indistinguishable from a human composition.
Similarly, the software Affectiva is able to deduce a person’s emotional state by examining their facial expressions on webcam videos.
As for marketing professionals, it seems many are convinced by algorithms’ capacity to produce imaginative content at least as effectively – and more efficiently – than humans. A 2017 White Paper by IDC, Can Machines be Creative? How Technology is Transforming Marketing Personalization and Relevance, reported that many marketers were actively thinking of using digital technologies to make creative content like images, color palette choices and taglines, as part of personalization campaigns.
Circling back to Jarno’s observation mentioned at the top of this article, it should be clarified that he wasn’t 100% critical towards AI-powered personalization. He did propose a solution; what he called ‘human-centered personalization’. In his words,
“Personalization should bring together collective intelligence and artificial intelligence. The connections become faster and the computers smarter and more efficient […] We can help each other to find and discover meaningful signals […] Human-centered personalization brings together human-curated signals and adaptive machine-learning solutions. In this way intelligent systems mature by learning from our individual and collective interactions and insights. In this way human imagination and irrationality can outplay the restrictions of algorithmic determinism.”
The idea here – which is perfectly applicable to today’s flourishing market for digital personalization solutions – is to have humans drive the tech, and have the tech augment humans’ capabilities. Or, using the words of our CEO Alix de Sagazan at our latest event, “Though we’re activity developing our platform’s artificial intelligence capabilities and recognize the advantages they can bring to our tool and our clients, we fully realize that human creativity is still the beating heart of our product vision. Predictive analytics can’t replace good human intuition – it should instead augment and facilitate its inception and development, from idea to test to implementation and analysis.”
This is the idea of intelligence augmentation – and it’s not a new one. Back in 1960, 5 years after the Dartmouth conference, (where it’s commonly acknowledged that the discipline of artificial intelligence was born), the psychologist and computer scientist J. C. R. Licklider wrote these words:
“Men will set the goals, formulate the hypotheses, determine the criteria, and perform the evaluations. Computing machines will do the routinizable work that must be done to prepare the way for insights and decisions in technical and scientific thinking. . . . The symbiotic partnership will perform intellectual operations much more effectively than man alone can perform them.” (J. C. R. Licklider, “Man-computer symbiosis,” IRE Transactions on Human Factors in Electronics, March 1960, http://worrydream.com/refs/Licklider%20-%20Man-Computer%20Symbiosis.pdf.)
If this vision is followed, ‘human-centered algorithms’ do indeed seem less ironic. If technologists use as their guiding light the idea that people provide the spark of ideas, and remain the ultimate focal point of an agenda facilitated by AI but kept in check by homo sapiens – then I think we can safely say that we’re making way for a ‘human-centered’ personalization all marketers can get behind.