Technologies tilt the balance of power in markets in particular because they change the sources of competitive advantage: some previous sources are no longer relevant and new ones appear. Artificial Intelligence is no exception to this history. How do sources of competitive advantage change with AI? Which are outdated? Which are new?
To answer these questions, Karkowski, Luger and Raisch analyse almost 20 years of chess game tournaments. During this period, three types of games were organised: conventional where individuals compete, centaur where individuals supported by algorithms compete and engine where machines programmed and fine-tuned by individuals compete.
The main findings of the study
- human capabilities are a significant source of competitive advantage in conventional chess (they take time to develop and cannot be easily imitated),
- chess engines can substitute human chess-playing capabilities rendering this resource obsolete (engines compute more quickly and are never tired or distracted),
- chess engines are not distinctive resources as a lot are available under an open-source license and leverage common technologies,
- humans can influence centaur and engine games outcomes (selecting the engine, tuning the parameters, …) and this influence leverages unique human creative capabilities. For example to exploit an opponent’s machine’s weakness or to select parameters that surprise their opponents. Humans can create new sources of advantage integrating the machine capabilities with their own capabilities. These new human-machine capabilities are valuable and are limited,
- the new human-machine capabilities are unrelated to human chess-playing capabilities. Players who win in centaur or engine games are not the best chess players but the good players who approach the game from a computational and data science point of view.
Implications
This study has two main conclusions. First, with AI, some previous human-based sources of competitive advantage disappear. However, as AI engines are widely available, they don’t constitute a source of a sustainable competitive advantage and it’s the combination of human and machine capabilities which are a strong source of competitive advantage. Second, in this combination of human and machine capabilities, data scientists and creative problem solver may perform better than domain experts.
Implications are numerous. On one side it means that companies need to develop human-machine capabilities through recruiting talents, training teams and developing hybrid teams of data science and domain experts. On the other side, from a competition perspective, it means the threat is no longer limited to companies within the industry but extends to companies from the tech and data industries which can leverage their specific expertise to enter non-digital markets.
Photo de Fredy Jacob sur Unsplash