When is an ‘AI’ Feature Actually Smart, and When is it Just a Buzzword?

Discover the true essence of AI in research kit and learn to differentiate between genuine artificial intelligence and simplistic automation merely labelled as AI for marketing allure.

The differentiation between artificial intelligence (AI) and oversimplified functions often labelled as AI is something that is harder and harder to understand, and something we see more and more in products. As labs increasingly adopt technology to enhance accuracy and efficiency, understanding this distinction helps in making informed decisions about integrating true AI into research processes and understands which products are worth looking at.

Understanding AI Beyond the Hype

AI, by definition, involves systems that can perform tasks that generally require human intelligence. This includes learning from data, solving complex problems, making decisions under uncertainty, and adapting to new situations without explicit programming for each scenario. For instance, an AI in a pharmaceutical research lab might analyse drug trial data to predict potential side effects not evident from initial observations, continually refining its predictions as more data becomes available.

The Misuse of ‘AI’ in Marketing

Contrast this with many products marketed as “AI-powered” which might only utilise basic error correction or simple conditional statements, like IF loops. Such features operate under predefined conditions, performing specific, programmed actions when certain criteria are met. For example, a storage system might be labelled as AI because it sends an alert when temperatures fall outside a set range—useful, yet hardly indicative of any advanced intelligence or learning capability.

This misuse of the term AI has become a marketing strategy to capitalise on the AI boom, often misleading consumers about the capabilities of the product. It’s essential for lab managers and researchers to scrutinise these claims critically.

‘True’ AI in Research Settings

In research settings, the expectations from AI technologies are significantly higher. True (a term we use loosely as the very definition of AI is problematic at times) AI systems are expected to handle complex datasets, adapt their functioning based on new information, and improve over time through machine learning. For example, an AI system designed for genetic research might learn from new genomic data to better predict disease susceptibility without being explicitly programmed for each genetic mutation it encounters.

This ability to learn and adapt is what separates true AI from simpler automated systems. It’s the difference between a system that can only execute predefined instructions and one that can devise new strategies to solve problems based on accumulating knowledge and data.

Evaluating AI Claims

When assessing a product’s AI claims, it is crucial to ask whether the system can:

  • Learn from its operations and evolve its responses.
  • Handle and interpret complex and varied datasets.
  • Make decisions in ambiguous or changing conditions without human input.

A positive answer to these questions typically indicates the presence of genuine AI capabilities, rather than simple automation labelled inaccurately as AI.

The Real Impact of AI in Laboratories

The impact of genuine AI is profound in research environments. AI can significantly accelerate research processes, reduce human error, and uncover insights that might be impossible to detect manually. For example, AI-driven analytical tools in bioinformatics can detect patterns in data that predict disease far more efficiently than traditional methods.

As laboratories continue to evolve into more advanced research hubs, distinguishing between real AI and simplistic, rule-based automation becomes paramount. The true power of AI lies in its capacity to think, learn, and adapt—qualities that go far beyond basic automation or error correction. Recognising when ‘AI’ is not just a buzzword, but a genuine feature, is critical for leveraging technology to its fullest potential in the pursuit of scientific advancement. It’s also important for cutting through the marketing spiel and finding products that can really advance your lab.


Matthew has been writing and cartooning since 2005 and working in science communication his whole career. Matthew has a BSc in Biochemistry and a PhD in Fibre Optic Molecular Sensors and has spent around 16 years working in research, 5 of which were in industry and 12 in the ever-wonderful academia.

Leave a Reply

Your email address will not be published. Required fields are marked *