Some insurance companies are using AI-based chatbots in customer service and continuously improving them through machine learning. Major companies like Bosch and Siemens as well as energy and utility companies are making big strides in predictive maintenance and field services. Additionally, financial advisors have developed robotic counterparts.
Conversations with software providers, consultancies and IT service providers show that while many companies are currently experimenting with AI technology, few of them have actually yielded positive results. A recent Luenendonk survey reported that 48 percent of IT service providers see an increasing number of projects concerning AI and RPA.
However, there’s still a number of factors which hinder AI innovation, such as a lack of AI experts and data scientists; legacy IT landscapes and therefore bad data quality; lack of enthusiasm or support from top management; or corporate culture and employee mentality being opposed to innovation in general.
Despite considerable obstacles, most companies are concerning themselves with artificial intelligence, either because they want to leverage competitive advantages or because they fear being left behind by a fast-changing market.
Europe is way behind China and the U.S. on AI efforts. Especially Germany, taunted as a hub for innovation, falls short of expectations. While the German government decided to invest three billion euros in AI until 2025, Shanghai is investing 15 billion dollars in AI projects.
However, what the government lacks in enthusiasm, German companies make up for with their efforts. For example, Bosch is planning to invest 35 million euros in a new AI Campus. In total, the company wants to invest 300 million euros in AI technologies until 2021.
AI is not possible without humans
According to aforementioned Luenendonk survey, a clear majority of companies (81 percent) is convinced that artificial intelligence will disrupt their industry. However, almost just as many (94 percent) think AI will continue to need some form of human supervision. Respondents think that fully automated solutions only make sense regarding more insignificant tasks involving high data volume, like automatically displaying adverts or giving product recommendations.
Companies want to use artificial intelligence to support employees, not replace them. For example, AI can help development teams verify and test their solutions or constructions; or it can help scan for faults in products through image recognition. Artificial intelligence can also be useful for the back office.
Regardless of which department they were talking about, respondents always highlighted that artificial intelligence is not supposed to replace workers or reduce personnel costs. Instead, AI tools are supposed to take care of monotonous, time-consuming tasks, so humans are free to focus on creativity and innovations.
Obstacles
A majority of respondents is convinced that artificial intelligence is going to be a game changer. However, most companies are currently only using AI in individual projects. It makes sense to be cautious; they are only just beginning to understand what AI is capable of and need first-hand experience before they can start transforming their processes.
Right now, implementing artificial intelligence is a trial-and-error process. In the beginning of an AI project, companies often don’t even know if the proposed strategy is feasible since required data is not available or they are unstructured.
Consequently, companies leverage an agile approach. Short feedback loops guarantee that the costs and efforts involved in a proof of concept stay within reasonable limits.
However, if data quality is lacking, no agile approach can save AI projects. Also problematic: most companies don’t have a standard definition of what artificial intelligence is.
Investments in AI
Willingness to invest in AI projects depends on the size of companies. The automobile industry is currently the frontrunner with budgets of more than 50 million euros. The reason is that autonomous driving is viewed as a future selling point.
In almost all other industries, budgets are way below ten million euros. Considering how omnipresent AI is in almost any discussion or even the news, these results are surprising.
Budgets are expected to increase significantly over the next years. Analysts strongly recommend increasing investments in research and development of AI. China and the U.S. are far ahead, but other countries can still catch up by investing in AI and other digital solutions like the IoT or autonomous driving.
Consultants claim AI will add $15.7 trillion in economic gains by 2030 while the reality is much less exciting. This type of hype is behind America’s mounting startup losses and disappointing IPOs, and the rest of the world should be careful to avoid it, as Mario so skillfully argues. My paper on technology and startup hype in Issues in Science & Technology, published by National Academies of Science, Technology and Medicine, describes the professional incentives and changes in online media that have encouraged hype about technology and startups to rise even as less comes out.
Jeff Funk
Retired Associate Professor National University of Singapore
Winner of NTT DoCoMo Mobile Science Award