It’s becoming impossible to ignore the advances and benefits of artificial intelligence in the workplace, which is why chief information officers already plan for its adoption — but what is the best way to approach implementing the technology?
In Gartner’s 2018 CIO Agenda, 3,160 CIOs were surveyed from 98 countries across major industries, including 461 government CIOs. It revealed 25 percent of respondents have already invested or are in short-term planning for AI.
AI does differ depending on sector, though. It ranked among the top 10 technology areas for the overall sample of those surveyed, but 19th for government. But in defense and intelligence, 7 percent of CIOs mentioned AI compared to the 6 percent of CIOs in other industries.
Regardless, the majority of CIOs said cybersecurity and AI will significantly change how they do their jobs in the near future, and CIOs increasingly adopt AI to boost customer experience or fight fraud. In fact, 4 percent of respondents have already implemented AI, and 46 percent developed plans to do so.
Yet, challenges remain, and though interest is high, implementation is still rather low. CIOs piloting AI programs face obstacles. And when planning for AI, 47 percent of CIOs said they needed new skills, and 37 percent found those skills hard to find.
So, for the nearly half of CIOs planning to deploy AI and for those who will inevitably begin doing so, where should they begin? Gartner analysts provided four lessons learned from early AI adopters:
Start Small and Aim Low: Don’t seek high outcomes or return of investments at first. Prepare for early AI projects to produce lessons learned that will help with larger pilots and implementations. Set a low financial target, and focus on small-scale adoption.
Augment People, Don’t Replace Them: Gartner predicts that by 2020, 20 percent of organizations will dedicate employees to monitoring and guiding neural networks. So, use AI as a way to engage workers with the tech and enable them to focus on higher-level tasks. AI can help humans make better, smarter decisions, and can be used to complete mundane tasks.
Prepare for New Skills and Talents: Most organizations lack the internal skills in data science (for mining and exploiting data) for thoughtful AI implementation, and rely on outsourcing that knowledge. Improperly preparing data for AI can cause bias in data, algorithms and AI systems. Prepare to store and manage large data sets for AI projects, and transfer knowledge from external experts to internal workers for long-term success.
Choose Transparent Solutions: Especially in government, AI projects will include systems and software from external service providers. It’s important to understand how that system is reaching decisions and why, which should be included in service agreements. This is particularly relevant when decisions are subject to regulation and auditing. If understanding the advanced analytical model is too complex, visualizations of the potential offers will do.