Effectiveness, Objectivity, and Ethics in Business Valuation Through Machine Learning Algorithms
Abstract: In today’s economic landscape, we are witnessing a progressive and unstoppable convergence between Artificial Intelligence (AI) and financial processes, opening new perspectives in business valuation. The integration of traditional methods (Discounted Cash Flow – DCF) with new algorithms (Machine Learning – ML) enables more dynamic and data-driven assessments, reducing the margin of error and offering better adaptability to the complexities of the current economic environment. However, the adoption of such technologies raises ethical concerns regarding transparency, reliability, and accountability in automated financial decision-making.
Keywords: #ArtificialIntelligence #GenerativeAI #Finance #Economics #AI #BusinessValuation #DiscountedCashFlow #TechInnovations #MachineLearning #DigitalTransformation #Ethics #Algoethics #DiegoSerafin #EthicaSocietas #EthicaSocietasJournal #ScientificJournal #HumanSciences #SocialSciences #EthicaSocietasSupplement
Lolita Guliman, psychologist and psychotherapist, university professor, expert in the assessment and treatment of distress among workers in high-risk sectors, and coordinator of working groups focused on supporting families, including through the use of cognitive artificial intelligence.
Diego Serafin, graduated in Statistical and Actuarial Sciences, with a specialization in Data Science, at the University of Trieste. He is a researcher in applied artificial intelligence, statistical modeling, and complex systems. He is passionate about data-science models applied to economic, financial, and business contexts.
Introduction
Artificial Intelligence (AI) is transforming the way we evaluate companies. Thanks to its ability to analyze large volumes of data and identify complex patterns, AI can help improve the accuracy of valuations. Furthermore, AI can reduce the subjectivity of assumptions and enhance the transparency of the valuation process. These tools, with their capacity to process massive datasets and learn iteratively, offer significant potential to improve the precision of financial forecasts and reduce the impact of human subjectivity. Business valuation is a critical activity for investors, analysts, and companies themselves.
However, integrating algorithmic tools into valuation processes is not without challenges. Ethical implications related to the opacity of models, potential biases embedded in data, and the accountability of automated decisions demand careful reflection on their use. In this context, it becomes essential to explore hybrid models capable of combining the theoretical rigor of classical methods with the flexibility and analytical power of emerging technologies.
The aim of this contribution is to propose a business valuation model that integrates the DCF method with Machine Learning techniques, highlighting its advantages, limitations, and implications from ethical, organizational, and systemic perspectives. While it does not analyze a real corporate case, the article presents a simplified illustrative example and focuses on the conceptual architecture of the model, as well as the critical considerations that its adoption entails for the theory and practice of business valuation.
The Traditional Model
The most widespread traditional method is the Discounted Cash Flow (DCF). For several decades, this method has been one of the most widely used and respected tools for estimating the intrinsic value of a company. Its strength lies in the underlying financial logic, which links a firm’s present value to its ability to generate future cash flows, discounted through a rate that reflects business risk. This approach is grounded in a solid theoretical framework consistent with the fundamental principles of corporate finance.
However, the practical application of DCF is far from neutral. The projected cash flows, the discount rate selected, the chosen time horizon, and the implicit growth assumptions all require a high degree of subjective judgment. The direct consequence is that valuations considered “objective” may, in reality, vary significantly depending on the assumptions made by the analyst. Furthermore, the DCF approach shows limited responsiveness to sudden changes in the economic and financial environment. Multi-year forecasts tend to crystallize expectations that may not materialize, especially in unstable markets, those with high volatility, or sectors undergoing rapid technological transformation.
Recommendations from international standards, such as the International Valuation Standards (IVS), clearly emphasize the need for prudence, transparency, and rigorous documentation in model choices and underlying assumptions. Nevertheless, even when correctly implemented, DCF remains a model highly dependent on the quality of historical data and the predictive capabilities of the analyst.
In summary, while the DCF method remains a fundamental point of reference, it exposes the valuation process to risks of distortion, inconsistency, and opacity—especially when not supported by tools capable of adapting to the continuous evolution of data and business contexts.
Innovative Models: Artificial Intelligence in Business Valuation
Artificial Intelligence (AI), and particularly the subset of Machine Learning (ML), is radically transforming the way organizations analyze data and make decisions. Even in the field of business valuation, these technologies are emerging as promising tools that can support—and in some cases redefine—traditional methods.
Unlike static models such as DCF, ML-based systems can learn from historical data, identify recurring patterns, and generate adaptive forecasts. Algorithms such as artificial neural networks (ANNs) and gradient boosting models (e.g., XGBoost) can manage a multitude of variables—financial and non-financial—capturing complex or non-linear relationships that are difficult to detect through classical analysis.
One of the main advantages of AI is its ability to continuously update predictive models as new data becomes available, thereby increasing the responsiveness of the valuation process to market developments. Moreover, the use of data-driven models can help reduce analyst subjectivity, promoting greater methodological consistency.
However, the introduction of AI into valuation processes also raises relevant concerns. The so-called “algorithmic black box”—that is, the difficulty in understanding the internal workings of complex models—can undermine transparency and the verifiability of valuations. Additionally, ML models depend heavily on the quality and neutrality of training data, with the risk of replicating or amplifying pre-existing biases. Finally, the use of AI in business valuation requires new professional skills and a redefinition of decision-making roles: human judgment remains essential, but it must coexist with automated systems in an equilibrium that is still evolving.
In this context, AI should not be regarded as a substitute for traditional methods, but rather as a powerful complement capable of strengthening predictive foundations, provided its use is guided by principles of responsibility, transparency, and critical awareness.
The Hybrid Model
To overcome the limits of traditional methods without abandoning their theoretical soundness, the model proposed here integrates the Discounted Cash Flow (DCF) approach with Machine Learning (ML) techniques, adopting a logic of methodological complementarity. A hybrid model that combines traditional methods with AI may represent the valuation framework of the future. Such a model can help improve valuation accuracy, reduce the subjectivity of assumptions, and enhance the system’s ability to rapidly adapt to constantly evolving market conditions.
The core logic is that ML models are used to enhance the forecasts of key financial variables that feed into the DCF, particularly future operating cash flows. Instead of relying on subjective estimates or linear projections, ML learns from historical, financial-statement, and—when available—exogenous data (industry, macroeconomic, sentiment data, etc.) to produce more robust and adaptive forecasts.
Once expected cash flows are estimated, the DCF method is applied as a financial valuation tool, discounting those flows based on a discount rate that can also be refined using predictive techniques, while maintaining human oversight of critical assumptions.
This synergy makes it possible to preserve the theoretical framework of DCF while strengthening its forecasting component and reducing the impact of subjectivity in assumptions. The result is a more dynamic, transparent model capable of rapidly adjusting to its environment.
Table 1: Advantages, Limitations, and Precautions of the Hybrid Model
|
Main Advantages |
Limitations and Precautions |
| More Accurate Cash Flow Forecasts | Operational and Technical Complexity |
| Greater Adaptability to Change | Dependence on Data Quality and Availability |
| Reduction of Subjectivity | Need for Interpretability and Human Oversight |
| Improvement in the Documentability of Decisions |
An Example of Synthesis
An investment fund aims to assess the growth potential of an emerging tech company, named TechKosei, currently developing a highly innovative product. Several sets of information are available to support the analysis: historical cash flows, sector data, information on competitors, and characteristics of the target market.
In this context, the adoption of a hybrid model unfolds in two main phases. In the first phase, a Machine Learning algorithm analyzes historical and contextual data to estimate the company’s future cash flows, taking into account market complexity and variability. In the second phase, these estimates are used within a Discounted Cash Flow (DCF) model to calculate the company’s present value.
As an example, while applying the DCF method alone—based on linear assumptions and deterministic scenarios—produced an enterprise value estimate of approximately 420 million dollars, the ML-enhanced integrated model generated a higher estimate of 500 million dollars, also highlighting a potential annual growth rate of 20% over the next five years. This difference stems from the predictive model’s ability to capture latent dynamics within the data and weak signals often overlooked by traditional methods.
The advantages of the hybrid approach are clear: stronger alignment with real data, adaptive capabilities, and the possibility of continuously updating the model as market conditions evolve. However, several critical issues remain, such as the availability and quality of data, as well as the limited interpretability of ML models, which require expert oversight to ensure ethically and methodologically sound use.
The adoption of hybrid models for business valuation, based on the integration of traditional techniques and Machine Learning algorithms, is not merely a technological advancement. It brings structural changes to decision-making processes, demanding critical reflection from ethical, organizational, and systemic perspectives. Four key dimensions emerge:
1. Transparency and Comprehensibility
One of the main risks associated with the use of AI in valuation processes is reduced transparency. Complex models—especially those based on neural networks—may appear opaque even to those who developed them. This algorithmic “black box” challenges the principle of traceability in economic-financial decision-making, hindering verification by third parties.
To ensure ethically sustainable use, it is essential to develop explainability mechanisms—tools and methods that make ML model results interpretable. Responsibility for the valuation cannot be fully delegated to the algorithm: human oversight remains indispensable.
2. Responsibility and Accountability
Automation of the valuation process raises questions regarding responsibility. Who is accountable for an incorrect valuation if predictions are generated by an algorithmic model? The company being evaluated, the analyst, the model developer, or the organization adopting the tool?
Organizations using hybrid models must therefore implement a clear accountability framework, distinguishing between human and algorithmic responsibilities, and defining internal policies governing their adoption.
3. Bias and Algorithmic Fairness
Another critical aspect concerns the potential reproduction of bias within the data. If datasets used to train models reflect preexisting inequalities or distortions, the model’s output may amplify these patterns. This can have significant implications for assessing creditworthiness, financial reliability, or growth potential, ultimately affecting investment decisions.
To mitigate such effects, action must be taken on data quality and representativeness, as well as on validation and continuous monitoring techniques.
4. Impact on Professional Roles and Organizational Structures
The adoption of hybrid models also reshapes professional roles within organizations. Today’s financial analysts are required not only to possess traditional expertise, but also a foundational understanding of predictive models, their limitations, and interpretation methods. This transformation requires companies to invest in training.
BIBLIOGRAPHY
- Arrieta, A. B., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., … Herrera, F. (2020). Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI. Information Fusion.
- Bartram, S. M., Branke, J., & Motahari, M. (2020). Artificial Intelligence in Asset Management. CFA Institute Research Foundation.
- Bhadamkar, A., & Bhattacharya, S. (2022). Tesla Inc. Stock Prediction using Sentiment Analysis. Australasian Accounting, Business and Finance Journal (AAB&FJ), 52–66.
- Bharadwaj, A., El Sawy, O. A., Pavlou, P. A., & Venkatraman, N. (2013, June). Digital Business Strategy: Toward a Next Generation of Insights. MIS Quarterly, 471–482.
- Brealey, R. A., Myers, S. C., & Allen, F. (2011). Principles of Corporate Finance. McGraw-Hill Education.
- Brundage, M., Avin, S., Wang, J., Belfield, H., Krueger, G., Hadfield, G., … Dafoe, A. (2020). Toward trustworthy AI development: Mechanisms for supporting verifiable claims.
- Chen, T., & Guestrin, C. (2016). XGBoost: A Scalable Tree Boosting System. KDD ’16: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 785–794.
- Copeland, T., Koller, T., & Murrin, J. (2000). Valuation: Measuring and Managing the Value of Companies. Wiley.
- Damodaran, A. (2012). Investment Valuation: Tools and Techniques for Determining the Value of Any Asset. Hoboken, Canada: Wiley.
- Enholm, I. M., Papagiannidis, E., Mikalef, P., & Krogstie, J. (2022). Artificial Intelligence and Business Value: a Literature Review. Information Systems Frontiers, 1709–1734.
- Fabozzi, F. J., Focardi, S. M., & Jonas, C. A. (2014). Artificial Intelligence in Asset Management. Wiley.
- Farahani, M. S. (2024, July 19). Analysis of business valuation models with AI emphasis. Sustainable Economies, 1–13.
- Floridi, L., & Cowls, J. (2019). A unified framework of five principles for AI in society. Harvard Data Science Review.
- Geertsema, P., & Lu, H. (2023). Relative Valuation with Machine Learning. Journal of Accounting Research, 329–376.
- Guidotti, R., Monreale, A., Turini, F., Pedreschi, D., Giannotti, F., & Ruggieri, S. (2018). A survey of methods for explaining black box models. ACM Computing Surveys, 1–42.
- Haich, A. P. (2021). Deep Learning Applied to Public Company Valuation for Value Investing. Thesis, North Dakota State University of Agriculture and Applied Science.
- Hitchner, J. R. (2011). Financial Valuation: Applications and Models. Hoboken, Canada: Wiley.
- Hoang, D., & Wiegratz, K. (2022). Machine Learning Methods in Finance: Recent Applications and Prospects. KIT Working Paper Series in Economics, No. 158.
- How AI is Transforming the Business Valuation Industry. (2024, April 28). Retrieved from Equitest – Valuation Methods Blog: https://equitest.net/business-valuation-blog/valuation-methods-blog/item/295-how-ai-is-transforming-the-business-valuation-industry.html
- IEEE. (2017). Cybersecurity and AI Ethics. IEEE Standards Association.
- IEEE. (2019). Ethically Aligned Design: A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems. IEEE Standards Association.
- International Valuation Standards Council (IVSC). (2024). International Valuation Standards (IVS). International Valuation Standards Council.
- Kantar, L., & Ayrancı, E. A. (2022). Estimating Financial Failure in Businesses Using Artificial Neural Networks: Turkish Manufacturing Industry Model Study. Journal of Corporate Governance, Insurance, and Risk Management, 327–340.
- Machová, V., & Vochozka, M. (2019). Analysis of Business Companies Based on Artificial Neural Networks. SHS Web of Conferences.
- Mehrabi, N., Morstatter, F., Saxena, N., & Lerman, K. (2021). A survey on bias and fairness in machine learning. ACM Computing Surveys, 1–35.
- Mestiri, S. (2024). Financial Applications of Machine Learning Using R Software. Munich Personal RePEc Archive (MPRA).
- Miciula, I., Kadlubek, M., & Stepien, P. (2020). Modern Methods of Business Valuation—Case Study and New Concepts. Sustainability.
- Mittelstadt, B. D., Allo, P., Taddeo, M., & Floridi, L. (2016). The ethics of algorithms: Mapping the debate. Big Data & Society.
- Nei, P. P., Tobi, S. M., & Jasimin, T. H. (2023). Big Data Application in Automated Valuation Model for Valuation Process. Journal of Advanced Research in Business, 1–15.
- Russell, S., & Norvig, P. (2021). Artificial Intelligence: A Modern Approach. Pearson.
- Torrez, J., Al-Jafari, M., & Juma’h, A. (2006). Corporate Valuation: a Literature Review. Inter Metro Business Journal, 2(2), 39–58.
- Tsai, H.-Y. (2021). The Impact of Artificial Intelligence on Sustainable Corporate Brand: a Netnography Study of Tesla.
- Unal Guner, P., & Unal, S. N. (2023). An artificial neural network-based method for company valuation. Journal of Business, Economics and Finance (JBEF), 91–101.
- Wilkowski, W., & Budzyński, T. (2006). Application of Artificial Neural Networks for Real Estate Valuation.
- Zugravu, A., Mansour, T., da Cunha, A., & Espérandieu, A. Using AI in economic development: Challenges and opportunities. Retrieved from McKinsey & Company: https://www.mckinsey.com/industries/public-sector/our-insights/using-ai-in-economic-development-challenges-and-opportunities/
LATEST CONTRIBUTIONS ON AI
THE IMPACT OF ARTIFICIAL INTELLIGENCE ON CORPORATE FINANCE
THE RISK OF RELIGIOUS BIAS IN ARTIFICIAL INTELLIGENCE
THE FUTURE OF FINANCE: THE IMPACT OF ARTIFICIAL INTELLIGENCE ON BUSINESS VALUATION
ARTIFICIAL INTELLIGENCE AND BIAS: LIMITS AND OPPORTUNITIES
GENERATIVE ARTIFICIAL INTELLIGENCE HOW DOES THOUGHT WILL CHANGE?
LATEST 5 CONTRIBUTIONS
THE ITALIAN REGULATION ON AI: NEW CRIMINAL OFFENCES AND ADMINISTRATIVE VIOLATIONS
“LIFE DOES NOT BELONG TO US”: MINISTER FLORES HERNÁNDEZ SPEAKS OF DIPLOMACY WITH A SOUL
TELEMEDICINE AS A PARADIGM OF TRANSFORMATION IN TERRITORIAL HEALTHCARE
THE MUNICIPALITY OF SANT’EGIDIO ALLA VIBRATA (TE) CONVICTED FOR MOBBING
LANGUAGE AS A TOOL OF VIOLENCE AND REDEMPTION
Ethica Societas is a free, non-profit review published by a social cooperative non.profit organization
Copyright Ethica Societas, Human&Social Science Review © 2025 by Ethica Societas UPLI onlus.
ISSN 2785-602X. Licensed under CC BY-NC 4.0