The Ghost In The Machine

I love this metaphor coined by Gilbert Ryle in the late 1940’s.

Despite the way it sounds, this phrase owes nothing to the title of a long-forgotten horror tale. Ryle coined it to try to popularise the extensive debate within certain circles at that time, as to where exactly the essence of human consciousness resides.

Is it purely within the physical body or is there something else?

For me though, it conjures up a powerful image of human thoughts and consciousness, as patterns of activity in the physical manifestation of our bodies. These patterns can be stimulated by events triggered both internally and externally, for example; hearing music, eating delicious food or re-running memories of our recent summer holidays – all of these create patterns of neural activity. These patterns of neural activity form our individual Ghosts.

As a species, we rely heavily on social learning and the behavioural patterns of our individual Ghosts are learned from our interactions with friends, family and wider social cultural values. Our perceptions of our cultural backdrop and our formative experiences will influence our individual value systems and how we respond to the events in our lives.

As people, our Machines contain the neural networks and the sensory nervous system typically described as the 5 Senses. Beyond this basic 5 though, there are many others not often mentioned. These include as examples, the sense of balance, the Vagus Nerve and the parasympathetic nervous system. The latter of these interestingly, has a significant influence on the control of our moods, immune responses, digestion, and heart rate.

Organisations’ Ghosts In Their Machines

In a previous article I argued that we can view our organisations through a zoomorphic lens. This means that we can also use Ryle’s metaphor as a powerful tool to visualise what happens at work and how we can make positive changes to it.

We know that our organisations create patterns of activity and data exchanges. Their manifestations of the Ghost are the triggers of, and responses to, events and the process definitions that we (hopefully) wrap around them. The External data that we use within our organisations for analytics and AI/ML correspond to the 5 Senses. They allow our organisations to percieve their ‘outside world’. The Internal data corresponds to the other ‘internal monitoring senses’ and can be used to indicate our organisation’s well-being.

For our organisations, the Machine in Gilbert’s metaphor corresponds to the physical infrastructure that supports our organisation’s behaviour. This includes the systems that we use and our working environments.

Notice though that our involvement in the activities of our organisations as individuals means that, at a certain level, we too can be viewed as ‘components’ of the Machine. Therefore, we can see that our individual behaviours will collectively have a major impact on the way that our organisations operate.

Thus, as a result of selecting people from similar backgrounds to form the core of our organisations’ functions, we will create a narrow base for our organisational responses, due to their narrow, shared base of experience and values.

This may of course be perfect.

If we are running a drug-use support line, it makes sense to select people who have a background that allows them to understand exactly the experiences of the people in need. This will make them better able to establish rapport and propose actions of which they have direct knowledge.

If our organisations are engaged in activities that have a wide-ranging effect on society though, we need to be sensitive to drawing our talent pool from people with a wide range of experiences and background. In my opinion, this also delivers the benefits of improving the creative strength of our orgainsations and also their resilience.

The AI/ML Ghost In The Machine

If we think about AI/ML and the potential for unintentional bias for a moment, we can immediately see the impact of Rye’s metaphor.

The algorithms we develop have intended bias and are trained to produce outcomes according to their requirements. But they are still created and trained by human input. As a result, the cultural bias of the creators will tend to influence the judgements of the ML and hence unintended or unconscious bias can be injected from the Ghosts of the creators. The choice of training data itself as an example, can reveal biases which are not easy for the team to spot as it reflects their own Ghosts’ biases.

If the creators come from a narrow set of backgrounds, then the chances of unintended bias can be exacerbated. To reduce this possibility, it becomes doubly important to ensure our organisations have cultural and experiential diversity of those working within the teams delivering AI/ML capabilities.

In addition, it is critical to make an honest appraisal of your organisation’s culture and value system. We have seen that the cutlural backdrop of your organisation will also impact the way that its analytics and AI/ML has overall beneficial outcomes for those inside the organisation, its consumer base and the wider society beyond.

In an era when Environmental, Social, and Governance (ESG) Criteria are becoming increasingly important, it makes sense to remember Gilbert Ryle’s simple phrase.

Leave a Reply

Your email address will not be published. Required fields are marked *