AI Meets... Stratecast’s Big Data & Analytics (BDA) Practice
Aug 07, 2015
I'm pleased to join my colleague in London, Lawrence Lundy, here on the AI Meets... blog, and happy to reassure him that AI won't be putting any competent researchers and analysts out of a job anytime soon. In case you missed it, I'm referring to his blog post here -- http://digitaltransformation.frost.com/ blog/ai-meets-research )
I’ll explain why our jobs are secure in a minute, but first let me give you a two sentence definition of AI:
• It’s nothing more (or less) than a growing collection of really powerful algorithms designed to find patterns and decide what to do about them.
• The decisions are based on logical inference, using statistical correlations and probabilities.
That’s it in a nutshell – whether the patterns are in speech, text, visual images, or machine data, AI’s machine learning algorithms can be taught (by humans) to find and use them. The current state of functional AI consists of rigid, purpose-built combinations of inference and pattern-recognition applied to very narrowly-defined use cases. An AI system that “learns” to play games (like Google’s Deep Mind) can’t recommend books; one that labels photos can’t translate English to French. In other words, no matter how smart today’s AI systems are, they can’t generalize their abilities to other applications.
This doesn’t mean AI systems will not find amazing new applications beyond enabling robots, self-driving cars and virtual assistants (like Cortana, Google Now and Siri). What it means is that bigger, broader AI applications will come with corresponding increases in the complexity of their parameters for pattern seeking, inference and optimization.
If someone launches Insight-As-A Service, as Lawrence suggests, it will be as a tool for human researchers and analysts, not a replacement. Lawrence and I should have our jobs for years to come, because data isn’t knowledge, computing isn’t thinking, and finding patterns isn’t the same as synthesizing information.
Readers who want to learn more, in plain English, about AI’s capabilities and limitations may want to check out my report, “Artificial Intelligence: A Practical Assessment,” published August 7, 2015, in our weekly series of Stratecast Perspectives and Insights for Executives (SPIEs).
Fatal error: Call to a member function getUserUserBio() on null in /home/frost/public_html/digitaltransformation/packages/problog/helpers/blogify.php on line 362