SAS study: AI is still in its infancy - despite the hype

SAS, one of the world's largest software manufacturers, presented the Enterprise AI Promise study today at the Analytics Experience 2017 in Amsterdam. According to the study, most companies have already started to discuss the potential of AI for their business or have launched their first projects. Overall, there is a great deal of optimism when it comes to AI. However, not all respondents are convinced that their company has the resources to exploit the potential.

The bottleneck here is not the technology - the majority are convinced that suitable solutions are generally already available. The far greater hurdles for most study participants are the lack of data science expertise, organizational reasons and a lack of social acceptance.

For 55% of respondents, the biggest challenge is the change in the world of work - i.e. the potential loss of jobs due to AI on the one hand and new job profiles related to AI on the other. Ethical factors were mentioned second most frequently.

Data scientists are missing

There is still a great need for action in terms of know-how within the company: Only 20 percent of the companies surveyed stated that their data science teams are well positioned for AI, while 19 percent do not yet have any such teams. Appropriate personnel programmes and organizational measures are planned by 28% of companies, while 32% want to tackle the development of AI knowledge through further training measures.

However, companies' confidence in AI is not yet very pronounced. Almost every second respondent (49%) lacks confidence in the results that AI delivers - this also applies in general to technologies that are a "black box" for most.

Only one in four companies has suitable infrastructures

The study also examined whether companies already have the necessary infrastructure in place. Here, a three-way split emerged between those who believe they and their infrastructure are equipped for AI (24%), the group who believe updates and changes are necessary (24%) and those who do not yet have a suitable platform for AI (29%).

"AI has made huge progress in recent times. Today, algorithms can perform tasks with impressive precision that were previously reserved for humans," explains Oliver Schabenberger, Executive Vice President and Chief Technology Officer at SAS. "Previously unthinkable: An algorithm beat the best Go player in the world. We had always thought that the game could not be played by a computer - at least not by human programmers. Now a machine has done it. Once the system had understood the rules, it learned to play - better than the best human player can. We can use this knowledge to solve business tasks better than today's static systems can. We can develop systems that learn business rules and then improve them on their own. That's what we're working on intensively at SAS."

For the study, responsible managers from 100 companies from the banking, insurance, manufacturing and retail sectors as well as representatives of public administration throughout Europe were interviewed.