Why improvisation is the future in an AI-dominated world
In his autobiography, Miles Davis complained that classical musicians were like robots.
He spoke from experience – he had studied classical music at Juilliard and recorded with classical musicians even after becoming a world famous jazz artist.
As a music professor at the University of Florida, which turns into a “AI University”I often think of Davis’ lyrics and how musicians have become more mechanical over the past century. At the same time, I see how machines are improving to mimic human improvisation, in all aspects of life.
I wonder what the limits of machine improvisation will be, and what human activities will survive the rise of intelligent machines.
The rise of mechanical improvisation
Machines have long excelled in activities involving the consistent reproduction of a stationary object – think identical Toyotas mass-produced in a factory.
More improvised activities are less rule-based, more fluid, chaotic or reactive, and are more process-oriented. AI has made great strides in this area.
Consider the following examples:
The trading stalls of Wall Street, Tokyo and London were once filled with the vibrant chaos of traders shouting and signaling orders, reacting in real time to changing conditions in a fluid manner. Most of these commercial pits replaced by algorithms.
Autonomous driving technology may soon replace human drivers, automating our fluid decision-making processes. Autonomous vehicles currently stumble where greater mastery of improvisation is required, such as deal with pedestrians.
Many live social interactions have been replaced by the sterile activity of carefully composing emails or social media posts. The predictive text of emails will continue to evolve, bringing an increasingly transactional quality to our relationships. (“Hey Siri, email Amanda and congratulate her on her promotion.”)
IBM’s computer Deep Blue defeated world chess champion Garry Kasparov in 1997, but it took 20 more years to The AI to defeat the best board game players is going. This is because the go has a much larger number of possible movement choices at any given time, and hardly any specific rules – it requires more improvisation. However, humans ultimately did not come close to the machine: in 2019, former world go champion Lee Sedol retired from professional gaming, citing the ascendancy of AI as the reason.
Music becomes more mechanical
Machines replace human improvisation at a time when classical music abandoned it.
Before the 20th century, almost all the great figures of Western artistic music excelled in composition, performance and improvisation. Johann Sebastian Bach was best known as an organist, with his first biographer describing his organ improvisations as “more devout, solemn, dignified and sublime” than his compositions.
But the twentieth century saw the bursting of the singer-songwriter-improviser tradition into specialized fields.
Performers were confronted with the rise of recording techniques which inundated consumers with versions of fixed, homogeneous and objectively correct compositions. Classical musicians constantly had to deliver technically flawless live performances to match, sometimes reducing the music to a sort of Olympiad.
Classical pianist Glenn Gould was both a source and a product of this state of affairs – he despised the rigidity and competitiveness of live performance and retired from the stage at the age of 31, but retired to the studio to assemble carefully visionary masterpieces impossible to perform in one take.
Most composers have abandoned the serious pursuit of improvisation or performance. Modernists have become increasingly fascinated with procedures, algorithms and mathematical models, reflecting contemporary technological developments. The ultra-complex compositions of High Modernism required mechanical precision on the part of the performers, but many postmodern minimalist scores also required robotic precision.
Improvisation ceased to be a part of classical music almost entirely, but flourished in a new art form: jazz. Yet jazz has struggled to achieve parity, especially in the United States, its home country, in large part because of systemic racism. The classical world even has its own version of the “one drop rule”: Works containing improvisation or written by jazz composers are often rejected as illegitimate by the classic establishment.
A recent New York Times item called on orchestras to open up to improvisation and to collaborate with jazz luminaries such as the saxophonist Roscoe Mitchell, who has composed numerous works for orchestra. But college and university music programs separated and marginalized jazz studies, leaving orchestral musicians deprived of training in improvisation. Instead, musicians in an orchestra are seated according to their objectively ranked ability, and their job is to replicate the movements of the lead player.
They are the machines of the music world. In the future, will they be the most disposable?
Davis perfects the art of imperfection
The march of AI continues, but will it ever be able to engage in real improvisation?
Machines easily reproduce objects, but improvisation is a process. In pure musical improvisation, there is no predetermined structure and no objectively correct performance.
And improvisation is not just instant composition; if that was the case, then the AI would make the distinction between the two disappear due to its computational speed.
On the contrary, improvisation has an elusive human quality resulting from the tension between skill and spontaneity. Machines will always be highly skilled, but will they one day be able to stop calculating and switch to an intuitive mode of creation, like a jazz musician moving from the rehearsal hall to the concert?
Davis reached a point at Juilliard where he had to decide his future. He was deeply connected with classical music and was known to walk around with Stravinsky sheet music in his pocket. He would later praise Bach’s composers in Stockhausen and record jazz interpretations of compositions by Manuel de Falla, Heitor Villa-Lobos and Joaquín Rodrigo.
Yet there were many reasons to abandon the classical world for jazz. Davis says he played “about two notes every 90 bars” in the orchestra. This contrasted sharply with the extraordinary challenge and stimulation of late night jam sessions with musicians like Monk Thelonious and Charlie parker .
He lived the reality of racism and “knew that no white symphony orchestra was going to hire [him]. “(By contrast, Davis regularly hired white players, like Lee Konitz, Bill Evans, and John McLaughlin.)
And he was the antithesis of a machine.
But in jazz, Davis has managed to transform his technical struggles with the trumpet into an iconic and haunting sound. Its wrong notes, missed notes and cracked notes have become hissings, whispers and sighs expressing the human condition. Not only did he have these “errors”, he also actively courted them with a risky approach that favored line color and expression over precision.
It was the art of imperfection, and this is the paradox of jazz. Davis left Juilliard after three semesters, but went on to become one of the most important musical figures of the 20th century.
Today the terrain has changed.
Juilliard has a thriving jazz program led by another trumpeter versed in both classical music and jazz – Wynton marsalis, who received two classic Grammy Awards for his solo work. And while the ‘robots coming for our jobs’ narrative is cliché, these shifts happen quickly, considerably accelerated by the COVID-19 pandemic.
We’re heading into a time when real robots could eventually replace Davis’ classic “robots” – perhaps some of the 20 violinists in a symphony orchestra – if only at first as a gimmick.
However, we may soon find that jazz artists are irreplaceable.
Rich Pellegrin is an Assistant Professor of Music Theory and an Affiliate Assistant Professor at the Center for Arts, Migration, and Entrepreneurship at the University of Florida.