Scaling AI: 5 things to consider

Principles & Frameworks
Ethics & Responsibility

Last month, I had the privilege of chairing the Scaling Intelligence Workshop at the AI Summit in New York. I gave a short keynote address emphasizing five important things to consider when scaling AI.

Scaling AI responsibly

Building and deploying AI models is not easy. These practical considerations could support a more successful endeavor:

  1. Scaling complex systems is hard, so be humble and realistic about the challenges and mistakes that are likely to occur. Don’t position AI as the obvious and guaranteed path.
  2. Realize these systems are not intelligent humans, so be sure to hold humans accountable for mistakes AI makes.
  3. Incorporate a multi-stakeholder development, deployment, and lifecycle governance approach with both technical and non-technical experts.
  4. Recognize the immaturity of operationalizing machine learning and the engineering talent often leading these efforts. Foster a supportive and collaborative atmosphere that complements your junior talent with people who have experienced when things go wrong or not as planned. 
  5. While AI might be new technology, many existing departments, skills, and experiences can contribute towards a successful outcome. Leverage existing teams and expertises as you approach new projects, but do not reinvent the wheel.

Holistic approaches to deploying AI

At Monitaur, we fundamentally believe in the power of AI to have a hugely positive impact on our daily lives. These five tips for scaling AI contribute towards a more responsible, ethical, and accountable project.

The tasks facing software developers and business executives are daunting. However, it is an enterprise’s responsibility to create the best possible environment for artificial intelligence to find success. To create this environment, there must be a sustainable approach to the development and deployment of AI, which can be achieved – in part – by considering these five tips.