AI Investment Threats: Beyond the Hype

This article highlights the investment risks associated with AI, including unintended consequences, technology vendor dependence, and cybersecurity threats, underscoring the necessity for thorough due diligence and ongoing risk monitoring in AI-related investments.

AI is not another technology trend

AI enables extremely fast product development. High computational power, cloud resources and freely accessible online data facilitate large-scale training and deployment of AI models. However, unlike other technologies that follow pre-defined algorithms and hard-coded instructions, AI can learn, adapt, and operate autonomously.

Private Equity firms are doubling down on their investments and AI technology integration into their systems. Unlike the dot-com boom and crypto hype, investors are facing much higher risks due to the widespread nature of this technology.

The Massachusetts Institute of Technology (MIT) maintains a repository of 1600+ AI risks, with varying likelihood and impact, and classifies how, when, and why these risks occur.

Some of these risks include:

Supply chain reliance:  Recent software products have been wrapped around mainstream Generative AI models like ChatGPT, giving rise to complex relationships with a lack of clarity and accountability over ownership, use and distribution of Intellectual Property, and its intrinsic value.

Asset depreciation: At the same time, as the technology degrades with use, it leads to asset depreciation over a much shorter time period compared to other technologies.

Concealed proliferation: AI is used in almost every organisation, big and small. It is embedded in Office 365, search engines and mobile phones. It is likely that tools such as report generators, data visualisation and data repositories now use AI. These features are reproduced in products without your company’s explicit consent and approval. AI is also used to construct and assess code bases.

Unintended consequences: While it is paramount that modern businesses continue to innovate and evolve, AI has a new set of risks and consequences, unlike all other types of technology.

Consequences for the investment industry

Investment risks are inherent in the financial markets and can significantly impact the performance of a portfolio. Understanding these risks is crucial for investors, as they can pose an existential threat to a portfolio company, potentially leading to a loss of investment. This overview will explore various types of investment risks, providing insight into their nature and implications for investors.

1.     Reputational risks

2.     Technological risks

3.     Systemic risks

4.     Regulatory risks

1.     Reputational risks

One of the most pressing risks of AI is the ethical implications surrounding its deployment. AI systems can inadvertently perpetuate biases present in their training data, leading to unfair treatment of individuals based on race, gender, or socioeconomic status. Furthermore, the lack of transparency in AI decision-making processes raises questions about accountability and responsibility. If an AI system makes a harmful decision, it can be challenging to determine who is liable - the developers, the users, or the AI itself.

2.     Security Threats

AI systems can also pose security risks, particularly in cybersecurity. AI is being used to attack at a fast pace, and also to thwart those attacks. Malicious actors can exploit AI to develop sophisticated cyberattacks that are hard to detect and counteract.

Phishing Attacks

AI can be used to automate phishing attacks, making them more convincing and difficult for individuals to recognise. Potential root causes can stem from training data, and solutions where one model performs based on the prompts of another model, thereby causing secondary level corruption.

Flo Health: Sensitive health data from millions using an AI-powered app finds its way to Facebook and Google. The Federal Trade Commission settlement compensates each person $1,000 USD - potentially amounting to billions.

Prompt Engineering

Simple prompts / queries can result in answers that have an existential impact. Further, it is not hard to construct prompts / queries that give users the answer they want and generate a false / incorrect response.

The National Eating Disorders Association (NEDA): Recommended routines that are dangerous for people struggling with eating disorders.

3. Supplier Dependence

As society becomes more reliant on AI-driven technologies, there is a risk of over-dependence, which can lead to vulnerabilities. Critical systems, such as healthcare, transportation, and finance, may become overly reliant on AI, making them susceptible to failures or malfunctions. In the event of a system breakdown, the consequences could be dire, affecting countless lives and disrupting essential services. Companies with a heavy, built-in reliance on foundational models can disappear overnight, due to technological advancements being released. 

Jasper.ai was too complex and too rigid to adapt to ChatGPT, resulting in a 42% decrease in market capitalisation.

OpenAI relies on AWS, which relies on NVIDIA which relies on Taiwan Semiconductor Manufacturing Company (TSMC). TSMC has a more diverse supply chain, with reliance on a relatively small number of silicon wafers, chemical and specialized equipment suppliers. Monopolisation of infrastructure creates system risks.

4.     Regulatory risks

The most recent regulatory guidance in Europe already makes CEOs accountable for unintended consequences, with cumulative fines that could total €55 million or 11% of annual turnover in the EU.

Therefore, investors need to be assured that their portfolio companies have comprehensive governance and an understanding of the proliferation of AI that navigates emerging regulatory risks.

"With the incoming EU AI regulatory framework, all business will need to be aware of the risks of AI" - Paul Lee-Simion, Chief Technologist, Ingenie

Critical Questions for Investors and company executives to Assess Strategic Risks

Identifying strategic risks in an investment portfolio is crucial for informed decision-making. Here are some critical questions investors should consider:

• How can one identify the unintended consequences of their investments?

• How can one determine what AI technologies are utilised within a company?

• How can one ascertain who is impacted by AI technologies – including developers, customers, and third parties?

Strategic Steps for Investors

While AI offers compelling opportunities for value creation, the risks and uncertainties it brings can scupper private equity investment into companies, especially when PE investors are unable to gain conviction about the technology, its defensibility, and its future. The black box nature of AI, the evolving regulatory climate, IP challenges, talent dependencies, scalability barriers, valuation bubbles, and reputational risks all combine to make AI-heavy investments more perilous than traditional deals.

For companies seeking private equity investment, understanding these concerns is critical. Transparency, robust governance, clear regulatory compliance, defensible IP, and realistic valuations are all essential to building investor confidence. For private equity firms, deep technical due diligence and a cautious approach to AI-related risks are now prerequisites for successful deal-making in the age of artificial intelligence.

Robust investment due diligence, as well as strategic and operating level governance must assure that AI is:

Operational level:

-       Not biased and is transparent and accurate

-       Ethical, reliable and beneficial

-       Subject to human oversight

-       Subject to continuous improvement and maintenance

Strategic level:

-       Conflicts of interest are managed, and output is interpreted.

-       Unintended consequences are understood and mitigated

-       C-suite review Internal Audits, manage risks and comply with regulations

The Ingenie Effect

Ingenie’s Finteco product is a forensic analysis of the existential risks to the investments private equity and others make in their portfolio companies. Ingenie’s Finteco product provides private equity firms with a comprehensive analysis of potential risks associated with their investments, helping them to identify and mitigate threats before they become problematic. By offering detailed insights into AI-related vulnerabilities and compliance issues, Finteco enables firms to make more informed investment decisions. This product ultimately enhances the ability of private equity firms to protect their portfolio value and achieve sustainable returns in an increasingly complex technological landscape. 

Using 21 categories across over a hundred risk contributors, Finteco quantifies the current state and presents strategic risk. In presenting strategic risks, Finteco uniquely combines the risks from contributors, that is, each contributor may be inconsequential on its own but existential when combined with others. Finteco then models the management of the risk, presenting the impact on each contributor with mitigation scenarios.

The holistic AI diagnostic / readiness framework – risk dimensions and the questions that must be asked.

Connect with our team to learn how Ingenie's Finteco Investment Intelligence Platform can empower your investment management strategy.

Schedule a meeting today to discover how we can help you quantify risks, optimise performance, and identify new opportunities across your investment portfolio.