How I Learned to Stop Worrying and Love AI
The current business cycle has been dominated by digital transformation. When we think about startups, they tend to be digital enterprises that benefit from a global customer base, cheap cloud-based software and low start-up investment. But like all business cycles, inevitably this one also has matured and is consolidating. There are now more than 1200 unicorns worldwide, while the combined market value of the top five tech giants (Meta, Apple, Microsoft, Amazon, Alphabet) is almost $10trn, more than one quarter of the entire S&P 500.
What comes next? Greater bandwidth (from 5G, and fibre), more means to generate data (IoT, social, cloud services), and new ways to process information (quantum computing, ml, blockchain), are setting the scene for a quantum leap in data management.
In his short story, The Library of Babel, the writer Jorge Luis Borges imagines an infinite library whose “shelves register all the possible combinations of the twenty-odd orthographical symbols.” How, he asks, can find meaning within this vast repository of information? “On some shelf in some hexagon (men reasoned) there must exist a book which is the formula and perfect compendium of all the rest: some librarian has gone through it and he is analogous to a god.” Now we have AI whose core value proposition is to be that universal interface which allows us to query our vast datasets.
The nuclear options
However, incorporating such a powerful new tool into business and everyday life will require a well-defined strategy, akin to military thinking around nuclear deterrence. In a market economy, as with the dynamics of nuclear adversaries, participants engage in a never-ending competition. Instead of aiming for a final resolution, every player’s objective is to sustain a presence in the game, thus perpetuating the business ecosystem.
Here are some core principles of how they may achieve this:
- Identify mission critical objectives and scenarios:
- Nuclear context: National nuclear doctrines clearly lay out the circumstances under which nuclear escalation is appropriate, typically setting policies for multi-tiered escalation and de-escalation paths.
- Business context: Clearly defined business processes should capture circumstances that require human intervention, decision and responsibility.
- AI adoption: AI may perform analysis and lay out evidence and projections at these junctions to facilitate human action.
- Centralise accountability for critical action
- Nuclear context: In most cases, only the head of the government can order (and therefore bear responsibility) for a nuclear strike. He or she may only do so if the pre-defined circumstances dictate such a course of action.
- Business context: Typically, the CEO and those with delegated authority may make business critical decisions which must meet logical, ethical and legal criteria.
- AI adoption: AI may manage and optimise adaptive business processes to execute the decision maker’s
- Verify and process action through dispersed accountability:
- Nuclear context: Nuclear escalation require evidence that predetermined conditions for escalation have been met. Any missile launch order must be independently verified in sequence by several layers within the command-control system, right up to the person who actually initiates the launch. Any one of these decision makers along this chain of command can block the launch.
- Business context: Businesses have multiple stakeholders who may independently interpret requests, each with the potential to refine or redirect an initiative of the CEO.
- AI adoption: AI may facilitate (operation and customer facing) communication, distributing information and aiding the business ecosystem in making independent verifications at all levels.
If the right processes are in place to keep us safe, AI might then achieve its promise of increasing productivity through higher automation, which is a key growth driver in our technically advanced and aging societies. As knowledge workers in particular get displaced, society may start to rethink how it redistributes wealth (for example, by expanding the bureaucracy or through a universal basic income). Access to resources may come to depend less on one’s financial status and increasingly on social factors such as our interests, identity or behaviour. Working life in one form or another will continue, but supported by ever more sophisticated IT tools.
In her book ‘The Human Condition’, Hannah Arendt distinguishes between labour, work, and action. Labour encompasses repetitive processes necessary for individual and collective survival such as cooking, cleaning and farming. Work involves purposeful creation for public consumption, such as writing articles or designing a product. Action revolves around political engagement, discourse, and deliberation, such as expressing an opinion. We might delegate the first two to machines. But ‘action’ is what most defines us as humans. If AI allows us to focus on this most human of activities, we may learn to stop worrying and learn to love AI.
Simon Barna is leading innovation, ai and digital transformation at BT.