AI can process data in seconds, but experienced professionals still provide the judgment that makes decisions trustworthy.
Over recent years, organizations have incorporated more technology than ever before, including data systems, automated reporting, and AI tools capable of summarizing information almost instantly. At SCS, we’ve developed tools over the last decade for measuring methane, visualizing underground conditions and equipment, designing more efficient renewable natural gas systems, and preventing pollution in buildings and other infrastructure. As with today, these advancements have reduced processes that once took days to mere minutes, improving operational efficiency, streamlining compliance and regulatory reviews, increasing safety, positively impacting operations, and public transparency.
Limitations of automated systems
While systems effectively report what is happening by quickly aggregating operational data, emissions trends, meteorological conditions, and historical patterns, they do not provide interpretation or judgment. Critical questions, such as whether data align with permit intent and regulatory interpretations, or whether conditions are isolated or indicative of larger issues, remain unanswered by technology alone and must be addressed by humans. Professionals with relevant experience and certifications are essential editors and reviewers for creating high-quality, trustworthy content for decision-making.
The shift in data interaction and its risks
The ease and speed of accessing summarized data have created a subtle distance from the data's origins. Previously, when data collection was more laborious, there was a deeper understanding of data sources, assumptions, and gaps. Now, users tend to trust summaries more than underlying processes, which can be risky if the summary misses nuances or anomalies.
From a regulatory perspective, this distance matters because a technically correct report might still raise questions, trends might mask deviations from assumptions, and systems indicating compliance might not withstand scrutiny. The legal responsibility remains with the business, which entails risk if AI-generated summaries are assumed to be infallible or are reviewed by less experienced staff.
Responsibility remains human
Technology can organize, highlight unusual data, and suggest areas needing attention, but it cannot assume responsibility. It does not sign engineering reports, respond to agencies, or explain decisions under challenge. This responsibility has not shifted with technological advancements; if anything, it has become more apparent.
Field experience and expertise matter
There are several examples of AI-enhanced environmental services that use advanced technologies and AI:
Example 1: SCS scientists and engineers can identify the geologic conditions under which an equilibrium approach underestimates critical pressure, as well as the operational details that affect pressure buildup in an injection zone over time. SCS professionals can identify conditions under which the equilibrium approach is not appropriate for evaluating the critical pressure, even though it is the current EPA standard. So, our professionals established a kinetic framework for its evaluation.
Our kinetic approach, using time-based single-phase computational modeling, is less complicated than the multiphase flow modeling already required by the Class VI regulations and is equally capable of calibration, monitoring, testing, and reevaluation during the operational phase of a Class VI project. In fact, EPA is currently reviewing the details of the SCS approach and considering revisions to its guidance.
Example 2: Hydrologic and Hydraulic (H&H) Modeling, used in engineering analyses to evaluate water bodies, pipes, culverts, channels, and rainfall. H&H Models enable the evaluation of drainage and/or flooding impacts of various development or restoration scenarios and support the development of design and policy solutions.
Challenges can arise when adeptness at locating and entering data exceeds the user’s experience in hydrologic or hydraulic studies. Balancing this knowledge is crucial, as both are essential for an accurate water resources study (including environmental, stormwater, and climate-change-related work).
The aspect of current progress that gives one pause is how easy it has become to get results with AI automation. While today’s users are adept at running simulations and getting results, sometimes these results are wrong or at least should raise questions.
A novice user may get results that look “fine” to them but are “odd” to an experienced water resource engineer. If not carefully reviewed for engineering judgment, the less experienced user could inadvertently issue plans or study results with costly errors. This is a critical reason for a seasoned modeling professional’s quality-assurance review. More formally, a project-specific QC human process, geared toward reviewing applied hydraulics in H&H modeling, helps maintain a client’s modeling performance quality and the documentation necessary for permitting, risk management, and agency communications.
Example 3: Drone methane monitoring is useful for identifying recurring issues, which significantly help balance a wellfield, thereby greatly enhancing methane capture and reducing air emissions. Landfills are complex, unique ecosystems that require infrastructure to continuously optimize the delicate interplay between gas extraction, moisture levels (leachate), temperature, and air intrusion to maximize safety, regulatory compliance, and renewable energy recovery. Collection of various systems data is optimized by tools such as SCSeTools® and SCS Remote Monitoring and Control®. Still, it takes a seasoned group of engineers, operators, and field technicians to use the results to design or improve landfill gas collection and control systems and to make informed decisions. Landfill engineers use the results to design landfills, extend cells, and operate the systems to improve their efficiency. The data analyses also supplement recycling and composting program experts who help divert what was once waste into new products or clean energy.
Tools are useful, but decisions are ours
Certainly, we affirm the usefulness of technological tools in improving processes and speeding data handling and modeling. However, we stress that interpreting information and making design and operational decisions remain human tasks that are not automated and affect many.
Faster data processing speeds up the process but does not simplify decision-making when designing a safe and sustainable environmental solution to prevent or capture emissions, increase operational productivity, or manage risk – the ultimate accountability lies with experienced people.
John Tsun is a project director at SCS Engineers. Diane Samuels is communications director at SCS Engineers.
