With all the hype surrounding artificial intelligence, it's easy for the average observer to overlook the costs and potential downsides to AI implementation in the energy industry. Despite a growing trend of AI adoption, each company must weigh the good with the bad.
These industry professionals pinpointed those challenges to AI adoption. Here's what they said.
1. Jason Kram, Executive Vice President at Adapt2 Solutions
“As with anything new, the primary challenge is that energy companies are slow to adopt and invest time and resources into AI. This is mostly due to a lack of understanding and education of the full potential AI has to offer. However, we are starting to see this shift as more enterprises leverage AI's capabilities to directly impact better trading strategies in front office market operations though not nearly fast enough playing catch up with leading industries like healthcare, finance, education and transportation.”
2. Greg Slater, General Manager at Flutura Decision Sciences and Analytics
“Companies are starting to think about AI adoption and the bigger picture—what can AI bring the company/ where can it be exploited to realize its potential to save time, resources and create ROI.Energy companies are thinking about optimizing their overall blueprint not about small problems for different departments, what direction that department should be headed in, or what sort of technology structure they need to realize the benefits. The focus is on utilizing AI to achieve the pre-determined digital strategy within your company. Energy companies are no different and need to articulate their blueprint internally before engaging with vendors to help solve their problems.”
3. Shuli Goodman, Executive Director of LF Energy
“One of the issues that utilities have is that most are quite locally or regionally focused and do not have the scale and scope to generate the necessary data- let's say – about transformers. This is an area where open data and shared observation could be tremendously helpful in being able to detect failure long before it happens. Open data could be incredibly important to accelerating AI – also most utilities do not the human capacity yet, so sharing and leveraging resources could help speed things up.”
4. Binu Parthan, Principal Consultant at Sustainable Energy Associates
“One of the major challenges is the limited data available to use AI effectively. Available data are scattered in silos and there is an opportunity to unlock data and integrate data and offer a much better opportunity to deploy AI.”
5. Jeff McGehee, Director of Engineering at Very
“Most people in the industry point to users and decision-makers as the blockers of AI adoption. This is true, but it’s not the root cause. Users and decision-makers will always be gatekeepers to new technologies, so the onus is on AI practitioners to provide AI solutions that meet the needs of users and decision-makers.
What is really holding back large scale AI adoption is the shortage of talented practitioners and managers that understand the nuances of bringing AI-driven technology into the real world. AI has shown promise as part of research initiatives in academia in industry, but there is a huge gap between the research in a lab setting and the messy, dirty, confusing world. When AI solutions provide reliable service, demonstrate measurable value, and can be delivered in an on-time and on-budget fashion, they will be adopted. The only people that can bring this to fruition are AI practitioners and their technical managers. We’ll have to wait for the talent pool to gain experience over the next 5-10 years before we see enough productive “AI engineering” teams to support industry needs.”
6. Larsh Johnson, CTO of Stem
“AI adoption is driven by the availability of a solid data and software platform that is able to present the required input data, both time series and otherwise, on demand at required resolution. Without this, AI prototypes can never become products. Stem's experience with 1000+ customers is a testament to this #1 need.”
7. Morgen Henderson, Journalist at Solar Power Authority
“Artificial intelligence isn't yet fully developed and requires experimentation to make it work for the intended purpose. It also requires mass amounts of data and internet access and sometimes multiple machines, meaning it uses a significant amount of power. Currently, AI utilizes many resources in order to work and is using much of the energy we're trying to conserve. This problem is currently being addressed and we hope it will be resolved in the near future.”
8. Edwin Chen, AI Solution Consultant at Dynam
“Entities that adopt and deploy AI solutions will need to invest the resources to accurately capture and exploit the data and at the same time ensure data integrity and security. As the supply and demand of energy becomes more automated, this process may be more suspectable and exposed to cyber-attacks. Under these circumstances, AI can be also used to counter and defend against these attacks.”
9. Mark Chung, co-founder and CEO of Verdigris
“Data. Right now, the physical landscape, supply, demand is all greenfield. There's the promise of a lot of IoT sensors but I think it's still a very cumbersome process of digital transformation/digital twinning. In order for that to take place, you need massive consumer adoption of sensors into physical environments. To achieve that adoption, we still need clear business values and business propositions. and significant capital to overcome the barriers in the market. Many of the applications for AI, have to be prefetched and hypothesized. I think market educations and significant available at-risk capital are going to be barriers.”
Have expert insights to add to this article?
Share your feedback and we'll consider adding it to the piece!ADD YOUR INSIGHTS