š¬ Research summary by Connor Wright, our Partnerships Manager.
[Original paper by Isabella Seebera , Eva Bittnerb , Robert O. Briggsc , Triparna de Vreeded , Gert-Jan de Vreeded, *, Aaron Elkinsc , Ronald Maiera , Alexander B. Merza , Sarah Oeste-ReiĆe , Nils Randrupf , Gerhard Schwabeg , Matthias Sƶllner]
Overview: The importance of collaboration in the AI space is not only between humans but also between humanity and AI. Imagining working with an AI teammate may no longer be imaginary in the future, and understanding how this will affect collaboration will be essential. For, understanding this will highlight the importance of the human cog to the human-AI machine.
Introduction
Have you ever imagined consulting an AI co-worker? What would you like them to be like? The implications of AI as a teammate are considered within this piece, stretching from how they look to how they could upset human team dynamics. While we must consider the benefits of this collaboration, the human element to the process must remain, especially in terms of human development itself.
Key Insights
A different kind of teammate
Whether the AI is in a physical robot or an algorithm, it cannot be compared to a regular human teammate. One key difference is its ability to assess millions of different alternatives and situations at a time, proving impossible for humans. While useful, the form in which the communication of this assessment arrives would need to be determined. It could be in speech or text, with or without facial expressions for visual feedback. Questions like these lead us to question what we prefer in an AI teammate over a human.
What do we want in an AI team member?
The paper holds the classic Alan Turing definition that āAI refers to the capability of a machine or computer to imitate intelligent human behaviour or thoughtā. In this sense, should our thinking about AI collaborators be centred in human terms? Like with chatbots, similar considerations are brought into play, such as whether the AI should have a gender, can it differentiate between serious and social chatter etc. Our decisions on these questions will then certainly affect how the team dynamic plays out.
The effect on collaboration
In this regard, it’s essential to differentiate between AI as a teammate and AI as an assistant. Collaboration with AI as a tool is not as thought-provoking as holding it in the āhigherā regard as a counterpart.
In this way, collaborating with such an entity could enhance or negatively impact the team dynamic. The AI could become a leader in the group on specific issues it can handle best, yet depending too much on the machine could lose human competencies. Furthermore, the AI teammate could prove excellent at drawing insights from data, but the lack of out-the-box thinking could reinforce already present views. Hence, while collaboration is undoubtedly affected by introducing AI, the right balance still needs to be struck to make the most of it.
Considerations when collaborating
Given the novelty of this practice and AI in general, why an AI would suggest a particular course of action becomes a critical question. In addition, the extent to which we recognise the AIās involvement can also have far-reaching impacts. Should the AI become a leader on a topic, should it be credited with its work? Much of this stems from whether AI can be creative or not, which can be found on poetry, fashion and music.
Between the lines
While collaboration with AI teammates may be essential practice in the future, I would be cautious against throwing such collaboration into every problem possible. Sure, using AIās analytical capabilities will nearly always be helpful, but that pertains more to AI as an assistant rather than a counterpart. Hence, problems such as trying to solve world hunger, I believe, would not benefit from an AI as a teammate intervention, mainly due to how AI can never actually feel or understand what being hungry feels like. Whatās for sure is that while AI collaboration can reap benefits, human involvement remains paramount.