The Turing Test is bad for business



Fears of the artificial Intelligence fills the news: job losses, inequality, discrimination, misinformation or even a superintelligence ruling the world. The only group that everyone thinks will benefit is the economy, but the data doesn’t seem to agree. In the midst of the hype, US companies have been slow to adopt the most advanced AI technologies, and there is little evidence that these technologies are a significant contributor to productivity growth or job creation.

This disappointing performance is not only due to the relative immaturity of AI technology. It also stems from a fundamental mismatch between the demands of business and the way AI is currently being conceived by many in the technology sector – a mismatch that originated in Alan Turing’s groundbreaking Imitation Game paper from 1950 and what is known as Turing -Test he suggested in it.

The Turing test defines machine intelligence by imagining a computer program that can imitate a person so successfully in an open text conversation that one cannot tell whether one is talking to a machine or a person.

At best, this was just a way to articulate machine intelligence. Turing himself and other technology pioneers such as Douglas Engelbart and Norbert Wiener knew that computers are most useful for business and society when they expand and complement human capabilities, not when they compete directly with us. Search engines, spreadsheets, and databases are good examples of such complementary forms of information technology. Although their impact on the business has been immense, they are not usually referred to as “AI” and in recent years the success stories they epitomize have been obscured by a longing for something “intelligent”. This longing is ill-defined, however, and with surprisingly little effort to develop an alternative vision, it increasingly means surpassing human performance in tasks such as seeing and speaking and in board games such as chess and Go. This framing has become dominant both in the public discussion and with regard to capital investments around AI.

Economists and other social scientists emphasize that intelligence not only or even mainly arises in individual people, but above all in collectives such as companies, markets, educational systems and cultures. Technology can play two key roles in supporting collective forms of intelligence. First, as highlighted in Douglas Engelbart’s groundbreaking research in the 1960s and the ensuing emergence of human-computer interaction, technology can improve the ability of individuals to participate in collectives by providing them with information, insight, and interactive tools . Second, technology can create new types of collectives. This latter possibility offers the greatest transformative potential. It provides an alternative framing for AI that has a significant impact on economic productivity and human wellbeing.

Companies are successful on a large scale when they successfully divide the work internally and bring diverse skills to teams that work together to develop new products and services. Markets are successful when they bring together diverse groups of participants and facilitate specialization to increase overall productivity and social well-being. This is exactly what Adam Smith understood more than two and a half centuries ago. Translating your message into the current debate, technology should focus on the game of complementarity, not the game of imitation.

We already have many examples of machines that increase productivity by performing tasks that complement those performed by humans. This includes the massive calculations that underpin the functioning of modern financial markets through to logistics, the transmission of high-fidelity images over great distances in the blink of an eye, and the sorting of tons of information to pull out relevant elements.

What is new today is that computers can do more than just execute lines of code written by a human programmer. Computers can learn from data and can now interact side by side with people, draw conclusions, and intervene in real-world problems. Rather than seeing this breakthrough as an opportunity to turn machines into silicon versions of humans, we should focus on how computers can use data and machine learning to create new types of markets, new services, and new ways to connect people economically rewarding ways.

An early example of such business-conscious machine learning is recommendation systems, an innovative form of data analysis introduced in the 1990s in consumer-centric companies like Amazon (“You Like It”) and Netflix (“Top Votes For You”). Recommendation systems are now ubiquitous and have a significant impact on productivity. They create value by using the collective wisdom of the crowd to connect individuals with products.

Emerging examples of this new paradigm include the use of machine learning to create direct connections between musicians and listeners, writers and readers, and game developers and gamers. Early innovators in this space include Airbnb, Uber, YouTube, and Shopify, and the term “creator economy” is being used as the trend gains momentum. An essential aspect of such collectives is that they are actually markets – economic value is related to the connections between the participants. There is a need for research on how machine learning, economics and sociology can be linked together so that these markets are healthy and generate sustainable income for participants.

Democratic institutions can also be supported and strengthened through this innovative use of machine learning. The Ministry of Digital Affairs in Taiwan has used statistical analysis and online participation to boost the kind of deliberative conversations that lead to effective team decision-making in the best-run companies.



Leave A Reply