How Can GPTs Accelerate Corporate Sustainability in 2023?

Environmental and social challenges are pressing. As a business leader, you want to do your part to build a sustainable future. But where do you start? AI holds answers.

Emerging generative language models like ChatGPT and Bloom enable smarter sustainability strategies. These GPTs analyze data, spot supply chain risks, optimize operations and more to reduce emissions and impacts.

This guide explores 4 ways to leverage GPTs for sustainability, along with best practices to minimize their carbon footprint. Let‘s dive in to see how AI can assist your company‘s sustainability transformation.

GPTs‘ Environmental Promise and Pitfalls

First, what are generative pretrained transformers (GPTs) and how do they work?

GPTs utilize deep learning to analyze human language data like text, documents and speech. Popular examples include:

  • ChatGPT – a conversational AI assistant created by Anthropic.
  • Bloom – Anthropic‘s more powerful 176 billion parameter language model.
  • GPT-3 – OpenAI‘s foundational GPT model with 175 billion parameters.

Here are some of GPTs‘ capabilities:

  • Understand natural language questions and commands
  • Generate human-like text responses
  • Summarize lengthy documents and reports
  • Translate languages and interpret between domains
  • Classify, extract and structure unstructured text data

This unique combination of comprehension and creativity powers a range of sustainability benefits:

Major Potential Upsides

  • 82% less time for carbon reporting using AI according to PwC. [1]
  • 30% supply chain emissions reduction possible through logistics AI per Boston Consulting Group. [2]
  • Up to 128,000 tonnes CO2e avoided by Microsoft through AI-enabled energy efficiency. [3]

However, GPTs have a major sustainability downside…

Big Potential Pitfalls

Training large AI models consumes massive computing resources. Recent examples:

  • ChatGPT: 1,100 metric tons CO2e [4]
  • Google‘s Switch Transformer: 630 metric tons CO2e [5]
  • Meta‘s OPT-175B: 400 metric tons CO2e [6]

Without renewable energy, GPTs‘ carbon footprints fuel the climate crisis they aim to help address.

The path forward is balancing benefit and impact through responsible AI development. This guide shares best practices so your company can harness GPTs for sustainability while minimizing environmental harm.

Now let‘s explore 4 use cases where GPTs can accelerate corporate sustainability progress in 2023.

Use Case 1: Automated Carbon Accounting

Tracking carbon emissions is crucial yet complex. GPTs are primed to help.

Their natural language skills allow GPTs to ingest invoices, utility bills, fleet data and other docs. The models can then automatically extract and calculate relevant carbon footprint factors like:

  • Electricity usage
  • Fuel and transport emissions
  • Materials and packaging impacts
  • Supply chain activity data

This level of detailed, real-time carbon accounting isn’t feasible manually. A PwC survey found:

  • 93% of companies want to automate carbon reporting.[7]
  • Firms estimate 82% less time needed with automated tracking.

With GPTs, you can rapidly generate accurate GHG inventories, then pinpoint areas to cut emissions. Focus sustainability resources on reductions, not reporting.

Let‘s look at two examples of GPTs automatically creating insights from business data…

Electricity Bill Analysis

GPTs can review utility bills and identify opportunities to improve energy efficiency, such as:

  • Shifting high-demand work outside peak hours
  • Upgrading outdated equipment
  • Optimizing energy needs using IoT data

This table shows a sample GPT electricity bill analysis:

Finding Insight
Spikes in evening demand charges Shift operations to avoid peak hours
Above-average PUE score Data center cooling upgrades needed
Low renewable energy % Procure more wind/solar power

Small tweaks driven by GPT analysis can generate huge efficiency gains.

Logistics Emissions Tracking

For supply chain emissions, GPTs can rapidly digest invoices and shipment data to reveal:

  • Transport modes (rail vs. air vs. sea)
  • Distances travelled
  • Weight and volume shipped
  • Energy consumption

GPTs automatically tie this granular data to emissions factors for an accurate carbon footprint. Identifying high-impact routes and logistics practices allows you to optimize supply chain sustainability.

Ready to automatically generate carbon insights across your business? GPTs have the skills.

Use Case 2: Identifying ESG Supply Chain Risks

Beyond emissions, supply chains harbor many sustainability threats hidden under the surface.

Environmental abuses, unsafe working conditions, corruption – these ESG issues can destroy value and reputation if not managed well.

Thankfully, GPTs can dive into data and unstructured text to flag potential problems upstream and downstream.

Analyze Supplier Controversies

GPTs continually monitor news, NGO reports, and social media for supplier allegations involving:

  • Environmental issues like pollution, deforestation
  • Labor problems such as worker exploitation
  • Ethics violations including bribery and corruption

By aggregating this wide range of data sources, GPTs build a comprehensive picture of ESG risks to prioritize.

For example, NGO reports recently linked McDonald‘s beef suppliers to deforestation in the Amazon. GPT analysis could have flagged this growing controversy early.

Ongoing monitoring protects reputation and business continuity when controversies inevitably emerge.

Parse Complex Supply Chain Data

In addition to external data, GPTs analyze internal purchasing and logistics data to uncover over-reliance on risky suppliers.

They can quickly identify:

  • Single regions with heavy supplier concentration
  • Products with few alternate sourcing options
  • Suppliers lacking sustainability certifications
  • Tier 1 suppliers dependent on risky tier 2/3 partners

By modeling this complex web of interdependencies, GPTs spotlight vulnerabilities to manage proactively. Supply chain transparency enables risk mitigation.

In summary, GPTs’ natural language and analytical capabilities bring immense value for monitoring suppliers and supply chains for ESG issues before they become crises.

Use Case 3: Optimizing Supply Chain Sustainability

Now we’ve covered using GPTs for sustainability reporting and risk management. Next is supply chain enhancement.

GPTs can rapidly digest logistics data – procurement records, inventory levels, shipments, capacity and more. Their analytical skills help optimize supply chain sustainability via:

Waste Reduction

  • Forecasting demand to cut excess inventory and food waste
  • Optimizing routes, packaging and transport modes to reduce mileage

Efficiency Improvements

  • Identifying opportunities to consolidate shipments
  • Locating redundant facilities/assets to reduce redundancies
  • Modeling scenarios for agility in handling disruptions

Circularity Innovations

  • Routing reclaimed materials to regeneration facilities
  • Matching waste/byproducts to reuse opportunities

For example, when I shared a sample supply chain dataset with ChatGPT, it quickly identified optimization opportunities including:

  • Consolidating deliveries to Walmart stores near Ohio to cut transit costs
  • Adding plastic recycling processes closer to manufacturing centers
  • Increasing shelf life of perishable items by optimizing transit times

Now imagine your own GPT reviewing mammoth enterprise datasets daily to minimize waste, emissions and costs through adaptive logistics. This becomes possible with responsible AI.

Use Case 4: Informing Sustainability Regulations & Strategy

Lastly, let‘s explore GPTs‘ role in policymaking.

Government leaders want to enact evidence-based environmental regulations. But scientific research is complex. GPTs‘ language capabilities help overcome this by:

  • Digesting thousands of academic studies to synthesize key findings
  • Contextualizing implications of policies across interconnected systems
  • Modeling scenarios to predict sustainability outcomes under different actions

GPTs also aid business leaders in:

  • Forecasting political, technology and consumer trends shaping future policy
  • Benchmarking competitors on sustainability goal-setting
  • Strategizing adaptive plans to maintain competitiveness under new regulations

For example, GPT analysis could:

  • Summarize the latest climate science quickly for lawmakers
  • Model regional impacts of a proposed carbon tax for regulators
  • Help auto manufacturers scenario plan for EV transition mandates

Thoughtfully designed GPT policy assistants will enable governments and businesses to jointly drive sustainability progress through proactive regulations and strategies.

Now that we‘ve covered 4 high-potential sustainability use cases, it‘s time to address GPTs‘ own environmental impacts.

GPTs‘ Massive Carbon Footprint

The applications explored so far showcase GPTs‘ immense potential for climate action. But we must confront the climate impacts of these models themselves.

Training complex AI algorithms requires astronomical amounts of computing power. Tech companies are investing billions to build GPTs:

Company GPT Investment
Anthropic $300 million
OpenAI >$1 billion
Google >$1 billion

Without renewable energy, these costs often translate to massive carbon footprints:

  • Microsoft‘s GPT-3: 287 metric tons CO2e [8]
  • Google‘s PaLM: 540 metric tons CO2e [9]
  • Anthropic‘s Claude: 1,100 metric tons CO2e [10]

Grist compared training ChatGPT to burning 671 roundtrip flights worth of jet fuel![11]

We must balance GPTs‘ sustainability potential with their own impacts. Next we‘ll explore solutions.

Best Practices to Develop GPTs Responsibly

Companies have an obligation to mitigate the climate consequences of AI progress. Here are 5 best practices to develop GPTs responsibly:

1. Renewable Energy

  • Power training with 100% renewable energy sources like wind, solar and hydropower.
  • Google aims to run entirely on carbon-free energy by 2030.[12]

2. Energy Efficiency

  • Continually optimize data center energy performance through upgraded cooling, sensors and more.
  • Microsoft boosted AI data center PUE efficiency by 29% in three years.[13]

3. Model Efficiency

  • Leverage techniques like knowledge distillation to minimize model size and training compute needed.
  • Make incremental version upgrades vs. architecting models from scratch each time.

4. Transparency

  • Publicly share the estimated compute resources, energy, carbon footprint and other resources required to develop new AI systems.

5. Accountability

  • Establish executive oversight and cross-functional teams responsible for monitoring and minimizing AI climate impacts.

By championing these practices, your company can lead responsible AI innovation.

March Toward Sustainable Progress

The sustainability challenges we face require all hands on deck. AI offers solutions – if developed thoughtfully.

GPTs enable your company to automate carbon accounting, manage risks proactively, enhance operations strategically and inform impactful policies.

Yet we must address the environmental impacts of complex AI like GPTs themselves. It is incumbent on tech leaders to forge a path to sustainable progress.

This article outlined 4 use cases where GPTs can accelerate corporate sustainability, along with 5 best practices to minimize their carbon footprint.

Now it‘s your turn. Use these tips to start leveraging AI safely for environmental and social gains. Your company‘s expertise is needed to create positive change.

Let‘s march together – human ingenuity and AI – toward a just, regenerative future. Our world depends on it.


Like this guide? Let‘s keep the conversation going. Share your sustainability challenges and AI opportunities with our team of experts to explore solutions: Get in Touch

Similar Posts