The outcomes of a global study involving over 2,800 technology experts on the realities of conceptual AI within businesses have been unveiled by O’Reilly, a reputable source for knowledge-driven learning in the realms of business and technology. The “Generative AI in the Organization” report delves into the utilization of relational AI by enterprises, the impediments to its adoption, and the expertise gaps that must be addressed to propel these technologies forward. The entire report is now available for free download.
As per Mary Treseler, the chief information officer at O’Reilly, “Generative AI presents a new realm of possibilities for businesses, enabling them to drive growth, streamline operations, and deliver exceptional customer experiences that differentiate them from competitors.” Nevertheless, this rapidly evolving technology could outpace business capabilities if the requisite skills are not in place to manage it effectively. The report underscores the critical need for companies to invest in skill development to stay ahead in the IoT revolution. We are still a long way from realizing the full potential of conceptual AI. It represents the best adoption of technology yet—and this is just the beginning.
A greater number of individuals have embraced conceptual AI compared to any other systems in recent memory. Over a third of respondents (37%) mentioned that their organizations have been leveraging AI for less than a year, while two-thirds (67%) indicated current usage of generative AI. Despite suggestions from entities like Gartner that AI may have reached the peak of its hype cycle, the survey findings suggest there is ample room for growth. The advancement of conceptual AI technology has made it easier to train models and develop intricate applications atop open-source models, utilized by 16% of respondents, requiring fewer resources to operate. The rapid succession of equipment generations is also accelerating the pace of implementation. Tools that automate complex prompts, facilitate archiving and indexing processes for modification, and employ vector databases for document retrieval are granting more organizations access to conceptual AI.
Challenges and Hurdles
Despite the widespread adoption of conceptual AI, many businesses are still in the nascent stages of implementation. While 18% of respondents claim to have programs in production, several obstacles hinder businesses from harnessing these technologies effectively. The top constraints cited by respondents include identifying suitable use cases (53%) and navigating legal, risk, and compliance considerations (38%).
The most sought-after skills include AI programming (66%), data analysis (59%), and AI/ML operations (54%). As the integration of generative AI accelerates, there is a growing demand for technology professionals with the expertise to advance these efforts. Respondents have realized the importance of general AI literacy (52%) when encountering the occasional anomalies displayed by generative AI tools.
When questioned about the risks associated with experimenting with AI, respondents highlighted concerns such as unexpected outcomes (49%), security vulnerabilities (48%), safety and reliability (46%), fairness, bias, and ethics (45%), and privacy (4%).
Artificial Intelligence in Action
Only 4% of AI users believe that generative AI tools will lead to a reduction in headcount, while 54% anticipate an enhancement in overall productivity. Currently, programming is the most prevalent application of generative AI (77%), utilizing tools like ChatGPT or GitHub Copilot. The top three use cases for generative AI in enterprises include data analysis (70%), customer-facing applications (65%), and aiding in content generation for marketing (47%) and other purposes (56%).
Other notable findings include:
- 34% of businesses are in the proof-of-concept phase with generative AI, indicating early-stage implementation across organizations. 10% are in the model-building phase, while 14% are engaged in product development. Furthermore, an impressive 18% have already deployed AI software.
- 64% of respondents have transitioned from off-the-shelf generative AI to developing custom applications, a significant progression necessitating financial and technological investments.
- While it is unsurprising that 23% of respondents utilize GPT models, 16% state that their organizations are expanding on open-source models, underscoring the continued importance and relevance of this field. The least popular models were LLaMA, Llama 2, and Google Bard (2.4%), respectively.
According to Mike Loukides, the vice chairman of information strategy at O’Reilly and the report’s author, “The adoption of relational AI is undeniably aggressive, but if we disregard the risks of premature adoption, we may face another ‘Artificial Winter.’” Enterprises have a significant opportunity ahead if they approach AI pragmatically, invest in training and tools, and think innovatively about its applications. The report concludes that while AI will not replace humans, businesses leveraging AI will augment human capabilities.
