Are most AI projects destined to fail?
Briefly

As businesses globally invest more in AI technologies, the expectations for productivity and efficiency growth remain high. However, many companies grapple with determining the effectiveness of their AI initiatives and the impact of project failures. James Hodge of Splunk emphasizes the necessity of shifting focus from viewing failed projects as mistakes to acknowledging them as learning experiences. He predicts that in the next few years, successful companies will heavily integrate AI, enabling them to adapt and thrive in their industries, highlighting the growing significance of domain-specific and smaller language models in AI development.
"I think there are 100% domain issues. One of the big predictions out there at the moment is where we're using lots of general, large language models, we're going to see a big shift towards domain specific language models, more reasoning models coming out, and even small language models, when we realize, 'Actually, I don't need to go and train something as large for the task I actually want to design it for.'"
"I think first realize it's not a failure, it's a learning experience. I think especially in industry, when we talk about project failure, it's the words 'failure', 'waste of money', 'no values returned', yes but actually what did we learn from that experience?"
"In three to five years, no business will be able to be operationally efficient and attractive to the market without the use of AI technologies. Because it will allow you speed to market, the ability to be able to be adaptive and responsive."
Read at ITPro
[
|
]