These Are the Top Challenges to GenAI Adoption According to AWS

(Roman Samborskyi/Shutterstock)

Businesses are making big bets on generative AI this year, according to a new survey on the nascent technology from AWS. While the investments are sizable, non-trivial challenges remain that have the potential to hamstring wider GenAI adoption, the company found.

GenAI rollouts are moving quickly, according to AWS’ new study, which is titled “AWS Study: Generative AI Adoption Index” and involved 3,739 IT decision makers across organizations in North America, Europe, and Asia. For instance, the survey found that GenAI was the number one budget priority for 45% of the IT decision makers surveyed, surpassing even security.

Source: AWS Study: Generative AI Adoption Index

Forty-four percent of survey respondents have moved beyond the proof-of-concept phase and either are working to transition their GenAI projects into production or actively integrating GenAI into their existing workflows, the survey found. One-quarter of the GenAI projects are still in the proof-of-concept stage, while 22% are in the exploratory phase. Only 8% have no adoption plans for GenAI.

The average organization has 45 GenAI projects or experiments in some stage, AWS found, with about 20 of them moving into production. They survey found 58% respondents are building custom GenAI applications on out-of-the-box AI models, while 55% are building custom applications on fine-tuned AI models using their own data. Forty percent are using out-of-the-box AI models as-is without any custom building or fine-tuning, while 25% are training their own AI model from scratch using their own data.

As organizations move forward with GenAI, they’re identifying problem areas and places that need further attention. According to the AWS survey, the top barriers including lack of a skilled workforce (cited by 55%); cost of development (48%); biases and hallucinations (40%); no compelling use case (19%); and lack of data (18%).

Source: AWS Study: Generative AI Adoption Index

More than half of the organizations AWS surveyed already have created training plans to upskill their workforce for GenAI, the survey found, while another 19% have plans to create one, meaning that three-quarters of all organizations surveyed will have AI training plans in place by the end of the year.

Nearly every organization (90% of those surveyed) say they’re looking to hire folks with AI skills. But nearly a quarter of organizations said at least half of their new AI hires “will need these skills,” the survey found.

However, AI upskilling may be easier said than done, as more than half of organizations told AWS that they don’t have a clear understanding of the AI skills their employees will need, while 47% say they lack the knowledge to implement a GenAI training program, and another 41% said they had a limited budget for GenAI training.

Related Items:

 GenAI Adoption: Show Me the Numbers

Global Survey By McKinsey: GenAI Adoption Starting To Deliver Value

KPMG Survey Reveals Strategic Shifts in AI Strategies in 2024