A recent survey by Appen, an AI data services company, reveals that the deployment of AI projects and their return on investment have declined. The survey, which included responses from 500 IT decision-makers across various U.S. industries, found that the mean percentage of AI projects deployed fell from 55.5 percent in 2021 to 47.4 percent in 2024. Additionally, the percentage of deployed projects that showed significant ROI dropped from 56.7 percent to 47.3 percent. Appen attributes this decline to a lack of high-quality training data, emphasizing that incorporating expertly labeled data can enhance the accuracy and relevance of AI models. Meanwhile, a Gartner report highlights that 49 percent of businesses struggle to estimate the value of AI projects, complicating adoption. Despite these challenges, EY reports that about a third of senior leaders see positive returns on AI investments, particularly in operational efficiency and customer satisfaction.
A recent Fedscoop report highlights significant challenges U.S. federal agencies face in adopting artificial intelligence. The report reveals that 29 agencies participated, with many citing data hurdles and a lack of AI-trained employees as major obstacles. Specifically, a dozen agencies noted issues with data management, while six reported insufficient funding for their AI initiatives. The Department of Energy expressed concerns over security issues with cloud services and a shortage of necessary graphics processing units. Furthermore, the Nuclear Regulatory Commission acknowledged a workforce that is both interested in AI and fearful due to a lack of understanding.
A new report from the Capgemini Research Institute highlights the transformative potential of generative AI in entry-level careers, predicting that it could facilitate 32% of entry-level tasks over the next year. The report indicates that generative AI tools may save junior employees an average of 18% of their time, leading to significant productivity improvements. However, despite 97% of leaders and managers experimenting with generative AI tools, only 15% use them daily. The research also suggests that new roles, such as data curators and AI ethics specialists, are expected to emerge, with 81% of leaders anticipating this shift. Furthermore, while many managers believe generative AI can enhance strategic decision-making, nearly half of teams use AI primarily as a tool rather than a collaborative partner. The report underscores the need for training and support to equip employees with essential generative AI skills, as only 16% feel adequately prepared for this transition.
These findings indicate that the federal government is struggling to keep pace with the AI revolution, which could hinder its ability to implement advanced technological solutions effectively. Federal agencies’ AI adoption struggles open consulting, training, and managed service opportunities, especially in securing infrastructure and training staff. AI adoption isn’t just about technology; it’s about cultural and operational preparedness.
Whether in the private or public sectors, AI’s success hinges on data quality. This calls for significant investment in better data management solutions and expertise. The Appen survey reveals a concerning drop in both AI project deployments and the returns they generate. The decline in deployment from 55.5% in 2021 to 47.4% in 2024 suggests that businesses are becoming more cautious about rolling out AI solutions.
It also signals that AI projects may be overhyped relative to their practical benefits unless this data challenge is addressed. To close this gap, the focus needs to shift to prioritizing data management and quality.