Since 2016, BTC Business Technology Consulting AG (short BTC) offers products and solutions for data analysis, AI and deep learning, which were developed by our AI team consisting of more than 10 experts. So far, we have carried out more than 20 projects on AI in the energy sector.

Based on our human-centric principle and approach, each solution offered by BTC has been aligned with various regulations such as laws, industry guidelines, conformities, and standards. In recent years, with the significant rapid development of AI, awareness of ethical and trustworthy AI has also increased significantly. Providing responsible and trustworthy products and solutions has therefore become one of our key strategies and commitment to our customers and partners.

In order to cope with the increasing use of renewable energy sources and their fluctuating and uncertain power generation, the power market needs to be flexible enough to ensure fast transactions motivated by short term forecasts of wind and photovoltaic systems. The ongoing revision of the intraday market creates incentives to use the full spectrum of the flexibility of the transmission system and should improve the coordination of power generation and consumption. But it is becoming increasingly difficult or impossible for human traders to obtain all the information and make reliable trading decisions in this complex context in less time. Thus, the need and importance of automatic and algorithmic tools and services for power trading has increased significantly.

But this new technology has also brought with it concerns and risks. One of the most common problems of many algorithmic trading providers is the lack of technical robustness and security resilience, which can lead to technical failures, such as bugs and overload. Other aspects such as auditability and explainability have also become critical challenges.

In order to follow the ethical guidelines of the AI HLEG proactively, we have selected a project from all the AI projects we have carried out for the evaluation of the TRUSTWORTHY AI ASSESSMENT LIST Within the project named "Algotrading", we designed and developed a software system to reliably predict the electricity price on the electricity exchange (EEX) and to generate bids suited to maximize the power trading profit for our customers. The various challenges such as uncertainty in power generation a highly dynamic and complex market, real-time response, high costs, etc. have been solved by the robust and comprehensive features of the system, while supporting the intervention of human traders. This system performs five main processing steps: 1) Collect market data, such as order book data, information about previous trades and general market KPI; 2) Store data in our company’s database; 3) Analyze and forecast data based on parameters; 4) Create orders, such as price and capacity to buy or sell; 5) Execute orders.

An outstanding feature of this system is its robust architecture, which meets various requirements, e.g.

• Resilience to attack

• Security by use of secure file transmission

• Fallback plan which ensures that the system only trades within normal price ranges and otherwise disables trading actions automatically.

Other features, such as a comprehensive communication mechanism, a fact-based trading strategy and explainable and verifiable algorithmic approaches, are also integrated into this system.

However, it would be more applicable if some parts of the TRUSTWORTHY AI ASSESSMENT LIST were tailored to the energy sector or even to the use case of power trading on the intraday market. But we at BTC continue to follow the guidelines and instructions based on our human-centric principle.