Here’s how organizations are commercializing artificial intelligence

 In |

Reading Time: 4 minutes

The continued evolution of artificial intelligence (AI) and related technologies is bringing new dimensions of computational functionality and value to business users.

Now more than ever, organizations should be looking to embrace these emerging technologies and learn how they might take full advantage of them.

AI is commonly categorized as one of three types: artificial narrow intelligence (ANI), artificial general intelligence (AGI), and artificial super intelligence (ASI). The distinctions among the three are a function of how well each mimics human intelligence and/or behavior.

A closer look at the 3 types of AI

ANI is becoming more common. Apple’s Siri and Tesla’s autopilot and computer vision code are examples of it.

AGI – also called Strong AI – is functionally equivalent to human cognitive ability in all aspects. Think of this like a machine that you could interact with in any way and believe it was human – including emotional responses and creativity. Researchers are still very far from this level of capability.

ASI is the most advanced kind of AI. This would be the dawn of the singularity, an AI that surpasses all things human, and has the ability to evolve itself. Most researchers and business leaders like Elon Musk and Mark Zuckerberg have stated publicly that they believe we are decades away from achieving ASI.

The availability and impact of commercial ANI

From a pure compute point of view, we have a long way to go.

In 2013, Markus Diesmann and Abigail Morrison were able to simulate 1 second of human, biological compute processing time with 82,944 processors and 1PB of memory over 40 minutes of real time. This was accomplished on the K computer at the Riken research institute in Kobe, Japan. That compute environment had 1% of the estimated nerve cells in the human brain.

However, ANI is currently commercially available and making an impact. An ANI is typically implemented to address use cases in predictive analytics, text to speech, image recognition, human-like chatbots, machine vision, natural language processing (NLP), expert systems, robotics, and planning or optimization problems. Very specific, well-defined use cases.

One ambitious project with ANI at its core is the Port of Los Angeles. It is the first automated port on the West Coast, and has deployed driverless, automated straddle carriers that work in conjunction with another automation system, Automated Stacking Cranes (ASCs). These immense, complex machines run ANI software to manage loading and unloading ships as well as loading shipping containers onto rail cars – all without human intervention.

ANI coordinates thousands of embedded magnets, lasers, sensors, safety triggers, and differential GPS to make this happen. One crane allows the port to redirect roughly 100 longshoremen – with an average annual pay of $200,000 – into other more critical jobs.

Vendors getting into the AI space

As with many new technologies, industry players want to be attached to emerging trends even if they don’t truly have an AI background. In the ANI space, we see many analytics companies jockeying for position. Organizations need to ask specific questions about the AI claims and capabilities that vendors offer.

The three major cloud vendors all have offerings in this space. Examples include AWS Alexa for Business, Azure Machine Learning, Google Cloud Machine Learning Engine, Amazon, Lex, Azure Computer Vision API, and Google’s TensorFLow.

The next steps for ANI

The next major step in capability for this level of AI is extending simple AIs into a broader interactive fabric.

In the same way simple, discrete applications hand off data across a messaging bus, AIs will be able to call each other and hand off processing functions and data across a common fabric.

Dr. Ben Goertzel and Cassio Pennachin have already established SingularityNET, a decentralized open marketplace for AI services. AI developers publish their services onto the SingularityNET network where they can be used by anyone with an internet connection. Participants can explore available AI services through a web UI. Google is working on a similar marketplace.

SHI too is bringing its expertise and thought leadership to the AI space with the development of an advanced chatbot system. Using largely open source tools, SHI has produced a chatbot AI with NLP/ML components that greatly reduces human interaction (reducing cost), provides a better overall customer experience, and reduces support tickets. Even better, the system learns in the literal sense, improving quality over time.

AI is also making an impact on the edge. By deploying AI to the edge in embedded systems, companies are able to use ever cheaper, lower power, and smaller edge devices. This delivers lower latency, conserves bandwidth, improves privacy, and most critically, enables companies to move smarter applications out to the edge.

Finally, the malicious use of AI is trending as a top concern among CSOs. Static scripts and largely manual tools applied to attack surfaces are no longer enough. Applying ML/AI to the business of cybercrime is emerging as a credible and serious threat. Today, attackers are actively using AI to bypass CAPTCHA systems, fine-tune phishing reconnaissance, and develop more aggressive, evasive forms of malware. These examples are just the beginning.

AI is going to continue to evolve. As such, organizations will need to stay up to date with this – as well as other emerging technologies – to ensure increased efficiency and deliver value to their end users.

To learn more about the commercialization of AI and how your organization can take advantage of it, contact your SHI account executive.