If you’re an avid fan of HBO's Silicon Valley, as I am, you may have recently downloaded the iOS app “Not Hotdog,” that was introduced on the show and developed in real life using real AI, in a ruse of ingenious marketing. The app is an image analyzing application leveraging Google's open-source TensorFlow to answer the age old question, “is this a picture of a hot dog?” The neural net model completes this task better than expected. There is a growing concern within IT leadership and business operators that not knowing how to engage with AI will put them in the rear view mirror, but in many ways, it's still early.
In the shorter term, these fears may be overblown. Outside of getting basic frameworks established, not yet on Stack Overflow or Github, there are hidden treasures in these and educational papers. This requires more research than most leaders might be used to, but it's where the best information can be found today. There's little doubt that the information available on the internet for developing deep learning applications will propagate exponentially in the future. For those who are willing to do the detective work, solutions are out there. The result is a deeper and broader understanding, as well as an overwhelming wonder at what's possible.
Some companies are using AI much more frequently in computer-to-computer activities and much less often to automate human activities. AI is used most frequently in detecting and mitigating computer security intrusions by IT departments. Even in this most popular use case, AI is not automating the jobs of IT security professionals out of existence. In fact, it’s helping often severely overloaded IT teams deal with the ever-increasing hacking attempts. In this way, AI is making IT security professionals even more valuable to their employers.
One of the fascinating parts of this rush to AI is the corresponding jump in engineers building their own computers with the express purpose of building deep learning models. It's a parting from the more standard use of cloud servers that form the backbone of traditional cloud based applications. Not that the resources to build these models aren't available in the cloud, however the required processing power is expensive.
There is a growing concern within IT leadership and business operators that not knowing how to engage some set of AI
This can quickly affect budget and if the model is not built correctly leads to extensive processing. Teaching a neural network is most efficiently done with a GPU, which can comprise of multiple of cores compared to a CPU.
One of the fascinating parts of this rush to AI is the corresponding jump in engineers building their own computers with the express purpose of building deep learning models. It's a parting from the more standard use of cloud servers that form the backbone of traditional cloud based applications. Not that the resources to build these models aren't available in the cloud, however the required processing power is expensive. This can quickly affect budget and if the model is not built correctly leads to extensive processing. Teaching a neural network is most efficiently done with a GPU, which can comprise of multiple of cores compared to a CPU.
IT is one of the largest adopters of AI, in computer-to-computer transactions such as in recommendation engines that suggest what a customer should buy next or when conducting online security risks. It isn’t just to detect attacks on the data center; IT is also using AI to resolve tech support issues. IT teams can automate the tedious work of new systems build outs, improvements into production, and insure staff is using approved technology.
There are very few companies using AI to eliminate jobs altogether. Where can your company find AI opportunities or applications of AI that won’t replace people and produce great results? Get AI implemented on processes that have an immediate impact on revenue and cost. Bring in the value of AI to decrease fraud, bad debt, and improve customer service. Leverage AI learning to make operations more efficient and effective. Find opportunities in which AI could help produce more products with the same number of people you have today.
Lead in with the back office, not the front office. You will see the greatest returns on AI in business functions that touch customers every day. Embedding AI in the products you sell to customers can produce immediate ROI, driving higher revenues and cost improvements. AI will have its biggest effect on the back-office functions of IT and finance. Conversely, AI’s impact on the front-office areas of marketing, sales, and services, take longer to develop. Focus AI initiatives in the back-office, predominantly where there are many computer-to-computer interactions like in IT and finance.
We know that computers today are far better at managing other computers digital information than they are at managing human interactions. Deploying AI in this way means jobs aren’t being eliminated and you as an IT leader can have an immediate impact on the efficiency and effectiveness of the business. Don’t fall into the job eliminating applications of AI that dominate the headlines. With the shift in responsibilities for the CIO, AI gives you the ability to reduce your daily workload and spend more time strategically driving the business.
Digital transformation and artificial intelligence will eliminate jobs that are for certain. With chatbots emerging in customer service roles and robots taking over the manufacturing floor, the future will bring job changes we can only imagine. However, IT leaders that wisely leverage AI in the areas where computers already interact have real opportunities to keep their teams engaged and highly valued for years to come.