Who can help? Carbon accounting options for busy executives

Regulatory tightening has spawned a host of new carbon accounting service providers offering businesses Carbon Accounting 2.0 solutions as Mandatory Era carbon reporting regulations kick in.

Most fall into one of the following categories:

  1. Legacy financial accounting services adding carbon reporting advice to their existing range of services.
  2. Start-up consultancies, offering to integrate a range of data services which include carbon reporting compliance.
  3. Sector-specific start-up consultancies, focusing on ESG compliance, carbon reporting, and accessing preferential financial markets in particularly energy-intensive industries, like steel or cement.
  4. Specialist materials databases offering access to LCA conversion factors along with technical and regulatory specifications.

Most are venture-capital funded data plays, highlighting the role AI can play in carbon reporting compliance.

Few modern businesses can afford not to include AI to deliver their services, but businesses are rightly becoming more sceptical about the technology’s claims. This scepticism is particularly justified when it comes to carbon reporting.

As businesses are rapidly learning, an AI solution is only as good as the data it’s trained on.

The kind of accurate carbon reporting now required by regulations like the EU’s Green Deal requires three broad data types:

  1. Scope 1 and Scope 2: data required to calculate their clients’ Direct and Indirect Emissions must come from their clients.
  2. Scope 3: data required to calculate Scope 3 must come from their SME supply chain.
  3. LCA: If reliable conversion factors are available for their clients, they will also need access to bespoke conversion factors, either for free from open source databases like EC3, created for the construction industry, or via paywalled commercial materials information databases.

To comply with Mandatory Era carbon reporting requirements, your carbon reporting partners need to satisfactorily explain how they will acquire this data – or how you can acquire it for them.

Rudiments of AI for busy executives

AI, it turns out, is not immune to the old computer adage of ​‘garbage in, garbage out’.

As the limitations and plausible hallucinations of Large Language Models (LLMs) like ChatGPT become more apparent, so do AI’s limitations in solving carbon reporting’s major challenges, in particular the SME Paradox.

To ​‘learn’, AIs need good teaching material.

If you train an AI on the Internet, it parrots back all the flaws, irrationality, prejudices and misinformation it finds there.

Train an AI on a high volume of high quality, consistently tagged, systematically taxonomised training data, and you’re likely to get highly accurate, useful, actionable results.

Nearly all the carbon reporting data collected to date is:

  • Hidden behind company servers or commercial paywalls
  • Generated using different methodologies
  • Those methodologies are themselves often hidden in IP-protected black boxes
  • Inaccessible without money.

This makes it either inaccessible to AI, or of such poor quality an AI is unlikely to be a useful tool. There’s currently little prospect of this changing, because:

  • Nearly all SMEs, who emit 70% of the world’s business emissions, can’t afford commercial carbon calculators.
  • In the absence of any legal compulsion, SMEs have no incentive to even use one of the thousands of free calculators available online, which vary significantly in their tagging and taxonomy.
  • Even if SMEs use one or more of the hundreds of free carbon footprints calculators available, their data ends up behind a paywall, usually for future commercial monetisation.
  • There is currently no carbon footprint equivalent of an open source, public LCA database, like EC3, to automate accurate Scope 3 estimates.

In short, carbon accounting 1.0 can’t deliver large-scale, high-quality carbon reporting because it’s ICOP:

  • Inaccurate
  • Costly
  • Opaque
  • Proprietary.

Logically, the solution should therefore be AFOT:

  • Accurate
  • Free
  • Open
  • Transparent.

This is the ecosystem niche See Through Carbon seeks to fill.