When chess grandmaster Garry Kasparov was defeated by IBM’s DeepBlue in 1997, he described himself as “the first knowledge worker to have his job threatened by AI.” At first he was angry, but within a year, Kasparov created Advanced Chess, combining machine processing power with human creativity to elevate the game to new heights. Nearly 30 years later chess players play alongside AI, proving Kasparov’s insight that machines can elevate human creativity rather than eliminate it.
At this year’s Sana conference in Stockholm, Kasparov joined AI luminaries including Wharton’s Ethan Mollick, Oxford physicist David Deutsch, OpenAI‘s Boaz Barak and Sana CEO Joel Hellermark – to explore how AI is reshaping work and human potential.
Rebuilding Organizations
Technologically the moment is ripe for transformation. “Organizations are at an inflection point where the breakthroughs in AI are starting to make stuff possible that wasn’t possible before,” according to Oscar Täckström, Chief Scientist at Sana. “But to really reap all these benefits, I think we really need to rethink how we’re doing some things and start to reimagine what could be done. And I think we have all the ingredients to do that.”
With machines increasingly handling repetitive tasks, Mollick emphasizes that organizations “have to be rebuilt from the ground-up”. This includes the need to redesign organizations for human-machine optimization, otherwise as Mollick points out “if you maximize for something [inappropriate]
, you’ll get the thing you maximized for.”
Incentivizing in a World of AI
Most companies continue focusing on efficiency and cost-cutting rather than leveraging human creativity. However, some are pioneering new approaches. Mollick shared: “I’ve talked to one company that gave out $10,000 cash prizes at the end of every week to whoever did the best job automating their job.” Such incentives, he notes, require leadership with growth intentions rather than a cost-cutting orientation.
The problem partly stems from our modern obsession with metrics. Sari Azout, founder of Sublime, observes: “If we go back to ancient Greece, human worth was really tied to wisdom and contemplation. Medieval England was incredibly tied to religious devotion. If we look at many indigenous cultures, status was deeply tied to spiritual connections, storytelling abilities, relationships. We just decided at some point that progress met numbers on a spreadsheet.”
Perhaps we should incentivize for wisdom instead. Mollick and others’ research reveals that “the only people who know how to use AI will be the people who are experts. [A second paper shows that]
junior people are much worse at using AI than senior people.”
This brings up a whole series of questions about employing junior staff, such as what value can junior staff bring if AI can fulfil the work better? How do we allow junior staff to gain experience when they add little value?
However, to re-define the use of humans in the workforce we need to understand how humans are valuable. In David Deutsch’s opinion: “I think people are valuable because they are different. Everybody is unfathomably different from everyone else. That fact is not being harnessed enough and can be harnessed more.”
Augmented or Replaced?
The debate about AI’s future largely hinges on whether humans will be augmented or replaced – and this may depend upon Artificial Super Intelligence. Oxford philosopher Nick Bostrom defines Artificial Super Intelligence as “any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest.” The key word is “virtually” – what remains are uniquely human capabilities that, when augmented by AI, create humans with enhanced abilities.
According to Barak, today’s AI intelligence “functions more like a beginner graduate student…. If we extrapolate they [AI] may well speed up science by being amazing research assistants, then later research collaborators and eventually independent researchers.” Anthropic CEO Dario Amodei characterized future AI systems as akin to a “country of geniuses in a datacenter.” Yet David Pfau from Google DeepMind cautions “if you’re talking truly superintelligent across all sort of human aspects, I don’t know that we’re ever going to see that” highlighting the distinction between superintelligence that complements human abilities and AGI that matches humans across all domains.
Azout suggests that with large language models, AI effectively becomes human collective intelligence. But we still come back to the need for wisdom. Like Kasparov, Azout sees the human role as guiding and evaluating AI outputs.
In Kasparov’s words: “We are not being replaced, we are being promoted.”
While these perspectives are encouraging, the prevalence of zero-sum thinking suggests the difficulty of enacting lasting change. Initiatives like Sana founding of the Swedish AI Reform foundation offer practical solutions. This non-profit, supported by Sweden’s Prime Minister, provides agentic AI free for two years to 2.3 million Swedes – including civil servants, students, teachers, researchers, and non-profits. By giving nearly 25% of Sweden’s population access to agentic AI, Sana is promoting the core message of its conference: humans must learn to work with AI, not against it.
I’m caught between seeing AI as a job-taker (even if temporarily) and as Kasparov suggests, a force that ‘promotes’ humans to higher-level work. But regardless of which view proves correct, one thing is clear: our organizations aren’t built for this shift. Companies are still optimizing for efficiency metrics from the pre-AI era rather than reimagining how humans and machines can work together. Unless we fundamentally rethink how our commercial and civic institutions operate, we’ll remain stuck measuring industrial-age outputs in a post-AI world.
