top of page

[Video on the homepage begins]

Video Title: Reclassify AI. Reshaping Data Intelligence Culture.

Artificial Intelligence is rapidly shaping our future.

AI has come into the mainstream because it allows us to make sense of an increasingly large amount of data in real-time.


80% of emerging tech has AI foundations.


75% of commercial enterprise apps use AI. 


Given all of the promise, a number of key deficiencies in the implementation of AI technologies remain.

Machines have no common sense.

Machine vision can recognize the objects in this picture, but can it tell us what is unusual about it?

Most AI models fail to contextualize information adequately. As a result, systems that may perform well for one problem cannot replicate that success for closely related problems.


Current Data Intelligence Culture:

  1. Highly Siloed Data

  2. Bottlenecks to Access

  3. Barriers to Expertise.

Algorithmic Bias Is Bad. Discovering Bias Is Good.


Bias Impedes:

  1. Facial Recognition

  2. Recommendation Engines

  3. Natural Language Processing

  4. and more…


Researchers from MIT and Stanford University found gender and skin-type bias in commercial facial-analysis software. The error rate for dark-skinned women was nearly 35 percent while only 1% for light-skinned men.

According to a 2019 study from the Georgia Institute of Technology, models used by self-driving cars, were 5% less accurate in detecting dark-skin versus light-skin.


Do Data Differently.


Reclassify AI. Reshaping Data Intelligence Culture.


What’s the Re in Reclassify?

  1. Semantic machine intelligence

  2. Implementors not just facilitators

  3. Diversity of thought is at the table

  4. Engender a culture of transparency

  5. Commitment to community.


Returning to our common sense example, The underlying problem: 
Today’s machine learning models merely learn statistical patterns without any real understanding of whether those patterns actually legitimately relate to the task.


This is deeply concerning because machine learning applications will be in a position to make life or death decisions, e.g., self-driving cars or military software.


Context matters. Humans use background knowledge and experiences to reason about interrelationships between subjects and objects. 


Semantic frameworks encode meaning and unlock understanding.


Semantic technology is improving AI by making systems more perceptive, intelligent and collaborative.


[End of video]


bottom of page