In today’s digital age, trusted information is the cornerstone of effective decision-making. It's also an essential ingredient in the adoption of artificial intelligence systems. But what exactly defines “trusted information”? Let’s start by breaking it down.
Data vs. Content: Understanding the Trusted Information Basics
Two main types of information exist: "Data" is well understood, while "Content" is less familiar. This concept can be simplified into a formula: Data + Content = Information.
Data refers to structured information, typically organized into rows and tables, such as databases. Content, on the other hand, is unstructured data - everything that doesn’t live in a database. This could include images, video clips, audio clips, social media posts, desktop documents, spreadsheets, PDFs, and more.
Structured data provides the who, what, where, and when of a situation, but it’s the unstructured data that reveals the how and why - often the most critical insights. Together, structured data and content give us a complete view, including the context, of any information use scenario.
The Collision of Data and Content in Business
If you’re reading this, you likely understand that businesses are evolving rapidly in how they manage, use and govern all types of information. This blog post is part of a series discussing information governance and the importance of trusting information in systems of record, including generative AI systems (artificial intelligence) .
You’ve probably seen statistics like 42% of managers admit to using the wrong information at least once a week. How does this happen in an era dominated by technology? Imagine a doctor making a life-saving decision based on outdated or incorrect treatment information - it’s absurd, right? Yet, this is precisely how many organizations handle their information assets, especially content. So, what’s missing?
The Missing Ingredient: Trust in Information
The key element we’re missing is trust. We need trusted information, delivered at the right time, to make accurate and informed decisions.
Considering the average information worker spends:
14.5 hours reading and answering emails,
13.3 hours creating documents,
9.6 hours searching for information,
9.5 hours analyzing information.
It’s clear that trusting the information we rely on is essential to productivity and efficiency. This is increasingly important with the onslaught of artificial intelligence systems being used in many decision making situations.
What Happens When We Don't Trust Our Information?
When we don’t trust the information we’re using, negative consequences will arise, especially in decision-making and business contexts:
Paralysis in Decision-Making: Without trust in the information,it’s natural to hesitate or delay important decisions, fearing the potential risks of acting on faulty or incomplete data.
Increased Risk of Poor Decisions: When we question the reliability of information, it becomes difficult to accurately assess situations, leading to misguided or suboptimal decisions. This can result in financial losses, inefficiency, or reputational damage.
Over Reliance on Personal Biases: In the absence of trustworthy information, people often fall back on their own assumptions, biases, or gut instincts, which are not always the best guide, especially in complex situations that require objective analysis.
Diminished Confidence and Morale: Teams and individuals may lose confidence in their work (or you) if they constantly question the information they are using. This can lead to low morale, lack of motivation, and diminished productivity.
Mistrust within Teams: If people feel that their information sources or leadership are unreliable, it can erode trust among colleagues and management. This can create friction, miscommunication, and a toxic work environment.
Wasted Resources: Mistrusting data may lead to unnecessary double-checking, verification efforts, or even discarding valuable resources that could otherwise be useful, wasting time, money, and effort.
Decreased Innovation and Risk-Taking: A lack of trust in information can lead to excessive caution, reducing the willingness to take risks or innovate, as teams may fear that new ideas are based on unreliable foundations.
In essence, trust in the information we use for decision-making is critical for confidence, efficiency, and strategic success.
Retaining Information is Also a Business and Legal Obligation
Legal, contractual and regulatory obligations also require the management of information as a "record" for record-keeping purposes. A "record" is recorded information that supports the activity of the business or organization that created it. A common definition is:
Recorded information, regardless of medium or characteristics, made or received by an organization that is evidence of its operations, and has value requiring its retention for a specific period of time.
Records managment is typically part of a dollars formal information governance program in most orgnanzations. Here is a link to ARMA (Association of Records Managers and Administrators) for additional information.
Defining Trusted Information: 5 Key Characteristics
So, what makes information “trusted”? Conceptually, trusted information is data or content that holds significant value and requires proper governance and retention.
Here are the five key characteristics of trusted information:
Authority: It is up-to-date and recognized as the authoritative, reference copy.
Authenticity: It is exactly what it claims to be and can be traced back to its original source.
Reliability: It accurately represents relevant facts, transactions, or processes.
Integrity: It is complete, unaltered, and maintains its context and chain of custody.
Usability: It is easily accessible, retrievable, and interpretable when needed.
Moreover, trusted information must be housed in trusted environments - such as secure repositories of record. If the environment isn’t reliable, the information itself cannot be trusted.
The Role of Trusted Information in AI Systems
As businesses increasingly adopt AI (artificial intelligence) and machine learning, the importance of trusted information becomes even more critical. AI systems rely on vast amounts of data to function properly. However, without trusted information, the very foundation of AI can become compromised, leading to poor decision-making, biased outcomes, and unreliable predictions.
Trusted information must serve at the core of effective AI systems:
Training Accuracy: AI algorithms need to be trained on accurate and reliable data. If the data is incomplete, outdated, or lacks authenticity, the AI system will learn incorrect patterns, resulting in flawed outputs. Trusted information ensures that AI is trained on the most relevant and up-to-date data.
Bias Prevention: One of the key challenges in AI is preventing bias. If the information feeding the AI system is not trustworthy or representative, the AI may reinforce existing biases. Trusted information ensures that diverse, reliable, and authentic data sources are used, mitigating the risk of biased results.
Improved Decision-Making: AI is often used to automate or enhance decision-making processes. By ensuring that only trusted information is used, businesses can be confident that the decisions or recommendations made by AI systems are based on accurate and comprehensive data.
Compliance, Privacy and Governance: As AI becomes more ingrained in business processes, maintaining compliance with regulatory standards like GDPR (EU General Data Protection Regulation) or industry-specific rules is crucial. Trusted information ensures that AI systems follow these guidelines, providing a level of transparency and governance that is essential in regulated industries like healthcare, finance, and legal sectors.
Data Security: Trusted information must be stored in secure, compliant environments. Ensuring that AI systems access information from these secure repositories not only builds trust but also protects sensitive data from breaches or misuse.
As organizations move towards a more information or data-driven future, integrating trusted information into AI systems is no longer optional - it’s essential for long-term success. Without it, the insights drawn from AI risk being unreliable, putting business operations and decisions at risk.
Business Imperative: Transparency of Information Used in AI Systems
It’s time for a new business imperative: full transparency of AI data sources. Right now, transparency is only partially available, and it’s not enough. Determining what data a generative AI system has been trained on remains difficult, as this information is often hidden under the guise of proprietary practices. While some organizations release high-level documentation, it’s far from a complete disclosure.
Hiding this information as a competitive advantage is a disservice to everyone, including the vendors themselves. We must demand full transparency into the data and training methods behind every system. This should be the minimum standard. Let’s push for this essential change!
What’s Next for Trusted Information?
In the coming posts, I will explore the broader topic of information governance and the implications of artificial intelligence on business decision making. But in the meantime, what do you think?
Do you agree with this definition of trusted information, or do you have a different perspective? Do you think we need transparency and trusted information in AI systems?
Kommentare