Expert Insights

In the film "The Cyborg Executive", Adam Ben-David, Senior Director of AI at Hyperexponential, uses the Android vs. Cyborg metaphor to illustrate two different ways AI can be integrated into software development. He challenges the widely hyped idea of fully autonomous AI (an Android approach), advocating instead a Cyborg model in which humans are at the centre of the system, controlling and guiding AI tools.

Adam argues that it is crucial to get comfortable with a Cyborg approach to AI in the early stages of AI development. Expecting AI to function flawlessly like an Android may be unrealistic and unproductive at this point. Instead, capitalizing on AI's ability to augment human effort can yield outsized returns.

Hear Adam elaborate:

  • The importance of keeping humans in the loop - especially in industries like insurance, where the potential for error is too high to let the system run blind.
  • The practicality of the Cyborg model, where AI boosts human capabilities - freeing up bandwidth for high-level work.
  • The two-part process that his team uses to generate deterministic code from probabilistic models – a technique that can potentially revolutionize the insurance industry.
  • The idea that software teams are essentially working in a Cyborg model using current development tools, and how they are looking for the highest leverage points as things start to blend.

Quote

quotation-marks icon
We are not replacing humans. We're moving toward something more akin to the 'Cyborg' model, where part of the work, the machine learning model-building work, is done by the AI. And part of it, the actual deployment, writing of the APIs to deliver the product, is still done by humans. So, it's more of a combination. The division of labour will be between what AI does best and what humans do best.quotation-marks icon

Monterail Team Analysis

Here's how teams can implement the Cyborg approach to AI-assisted software development:

  • Embrace the Cyborg over Android model: Recognize the value of integrating AI into workflows as tools that enhance human capabilities, as opposed to expecting fully autonomous systems.
  • Keep human judgment at the center: Make sure your AI system is designed to work with the valuable oversight and guidance of your team, especially in industries where the cost of errors is high.
  • Drive AI to boost higher-level work: Find opportunities to use AI to automate routine tasks, freeing your team to focus on more complex, impactful work.
  • Get comfortable with the Cyborg approach now: Building fluency with AI integration in these early days will yield significant gains and put your team at an advantageous position as AI progresses and becomes more prevalent.
  • Engineer for deterministic outcomes: As far as possible, design and train AI systems to generate deterministic outputs from probabilistic models to ensure the accuracy of automation processes.
  • Evolution of the workflow: Begin to view the AI system as a colleague who’s taking part in building the system, thereby identifying the highest leverage points for both human team members and AI agents.
  • Be creative: AI will not only help in dealing with the current workload, Adam also envisions AI opening up opportunities for underwriting new, unique risks – a lesson that can be extended to other industries as well.