Amid all the hype we’re hearing from the media and technology providers about “intelligent systems”, CIOs are still unsure of how they can apply artificial intelligence (AI) for long-term success.
A Forrester report guides CIOs on the steps they need to take to ensure their organisation gets the most out of the technology while minimising their exposure to risk.
“Sentient systems capable of true cognition remain a dream for the future. But many AI techniques and building blocks are available today, and you can leverage them as part of your process optimisation programs and customer facing applications,” Forrester principal analyst Martha Bennett writes in the report.
“Forrester surveyed business and technology professionals and found that, while only a comparatively small number of organisations have already implemented AI and are expanding their use, over half of companies expressed their intention to invest in AI in the next 12 months.”
According to Forrester Data, more than half (57%) of business and technology professionals expect that AI will help improve customer experience and support, with 37% implementing or planning to implement intelligent assistants and 35% doing the same with cognitive products for customers.
However, it’s clear that AI investments are still in the early stages and over half of survey respondents stated that they haad not seen any results of their AI initiatives.
That said, of those who did see some results, only 6% felt that results hadn’t met expectations.
Getting down to brass tacks, the report takes a close look at what is possible with AI today. Bennett writes that AI isn’t a single technology but rather an entire discipline encompassing a variety of different categories – some of which are ready for deployment, while others still have some way to go before they are ready for practical deployment.
The report goes on to caution against marketing puffery as vendors talk up ‘pure AI’ or true intelligence that is indistinguishable from (or even superior to) human intelligence. According to Forrester, some companies are bandying about “cognitive” functionality when, in reality, true cognitive computing remains in the realm of research.
Forrester has categorised various AI technologies that are ready for deployment in what it calls ‘pragmatic AI’. These include: speech recognition and natural language processing and generation; predictions using machine learning and knowledge engineering (for example, how Netflix optimises its recommendations through a variety of machine learning techniques); image recognition in combination with machine learning and deep learning; advanced discovery techniques; and robotics and self-driving cars.
Despite all it has to offer, AI implementations should proceed with caution. The report highlights that it’s easy to overlook just how much work goes into getting a computer system to the point required to deliver the promised results.
While it may vary from company to company, how much effort and time is required will depend on use cases and other variables including the availability of skills and usable data. In Bennett’s report, she warns CIOs to temper their expectations and be realistic about what is available out the box and to be realistic about tradeoffs.
The report also highlights the risks associated with AI which, ultimately, still depends on human input, training, and fine-tuning of the systems. Human intervention can include nefarious agendas of individuals and CIOs need to be on their guard. Moreover, AI systems can still behave unpredictably, and unsupervised decision making should be avoided. CIOs should be warned that they need to fully understand the risk and ethical conundrums that come with any AI implementation.
To ensure companies benefit from AI opportunities, but still mitigate the risks, Bennett advises CIOs to take responsibility for the organisation’s AI readiness.
Some of the specific advice to CIOs in the report include:
* Building a roadmap for AI deployments and ensuring that the correct skills are available to the company.* Taking an end-to-end process view, including how humans within the organisations will interact with the AI system.* Agreeing on ground rules and governance principles surrounding AI. This should happen at Board level and include discussions on how AI may affect jobs through change or even replacement.
Finally, Bennett’s report is emphatic that the CIO should place themselves at the helm of the AI strategy, taking responsibility and proactively driving the process.
“The greater the degree of unpredictability in an AI-powered system, the greater the likelihood that unforeseen negative outcomes will occur. That’s why humans — and their ability to reason — will remain an essential part of the equation for the foreseeable future, possibly forever.”