Top Enterprise AI trends: open the black box, address bias

big-data_data-scientist_artificial-intelligence_machine-learning-100747391-large

CIO magazine’s 2018 AI trends points out “most enterprises have yet to see quantifiable benefits from their investments in these areas”.  However, the level of investment in AI continues to accelerate, as the focus is shifting to include building an ecosystem around the core algorithms to provide the governance and trust needed by the enterprise.

As enterprises move forward with operationalizing AI, they will look for products and tools to automate, manage, and streamline the entire machine learning and deep learning life cycle …

With this push for greater trustworthiness comes increased focus on the black box problem and  biased training sets.

AI must solve the ‘black box’ problem … One of the big barriers to the adoption of AI, particularly in regulated industries, is the difficulty in showing exactly how an AI reached a decision … creating AI audit trails will be essential …

Bias in training data sets will continue to trouble AI … What has started to emerge as a key part of the conversation is how training data sets shape the behavior of these models …  models are only as good as the training data they use, and developing a representative, effective training data set is very challenging.

CIO magazine

These trends build the case for explainability and  XAI techniques.  Data scientist need to broaden our thinking to get past our current fallacies.

CIO magazine reflects the perspective of enterprise technology consumers.  Their emphasis on the black box problem and bias issues indicates those concerns are spreading from a specialist community to mainstream enterprise decision makers.

Leave a Reply