Using AI in Your Lab

Did you miss our free webinar?
Watch the recording now

Questions & Answers from the Live Event

ULT Freezers with Built-in Cloud-Based Temperature Monitoring Demo

Presented by:
• Jean Fallacara – MD, Serial Entrepreneur & Master of Chaos inTEST Corporation
• David Hayes – Global Product Manager Cole-Parmer, an Antylia Scientific company
• Cindy Gisler –Technical Specialist-Traceable® Cole-Parmer, an Antylia Scientific company

In this webinar you will:

  • The value of cloud-based data and AI to lab workers
  • The possibilities of this tech for the industry in the near future
  • Where VFC and COVID-19 programs’ vaccine requirements differ
  • How ULT freezers with TraceableLIVE cloud-based monitoring work into the strategy

As scientists and science enthusiasts, we often have a love/hate relationship with technology. On one hand it enables us to do things better, faster, or more nimbly. On the other hand it sometimes creates controversies, complications and difficult questions. The advent of revolutionary ideas like AI and cloud computing in the lab space is no different. Join us for a discussion among experts in the field, including leading AI technology serial entrepreneur Jean Fallacara, as they explore outcomes of forward-thinking tech in labs today.

This webinar is proudly presented by Traceable, the leading technology company for data monitoring temperature within research labs from biobanks to research labs and clinical trials and more.


Your presenters for this webinar

Jean Fallacara – MD, Serial Entrepreneur & Master of Chaos inTEST Corporation
Jean Fallacara, engineer in biotechnology with a specialization in virtual reality and augmented reality from MIT, is an expert in cold storage with integrated monitoring systems. Jean is currently managing director for inTEST Thermal Solutions after his R&D group Z-Sciences Corp was acquired in 2021. As an author, he is also known for writing "Neuroscience Calisthenics: Hijack your Body Clock".

Your presenters for this webinar

David Hayes – Global Product Manager Cole-Parmer, an Antylia Scientific company
David has over 10 years of experience with Cole-Parmer, an Antylia Scientific company and is currently managing the Cole-Parmer equipment line. His experience ranges from product concepts and development – sales and marketing – to working directly with customers.

Your presenters for this webinar

Cindy Gisler – Technical Specialist-Traceable® Cole-Parmer, an Antylia Scientific company
Cindy Gisler holds a BS in Medical Technology and a BA in Biology. For the last 21-years Cindy has helped design and market Traceable® monitoring products in the clinical and vaccine industries. She has extensive knowledge of the needs and concerns of these fields, including the importance and issues that arise with maintaining cold chain.

Questions & Answers from the live event

How do you know the information given back from ChatGPT is going to be reliable?

AI or artificial intelligence (ChatGPT) generates responses based on the context of the query. However, there could be times when it may not fully grasp nuanced or complex topics, which could affect the reliability of the responses. Moreover, ChatGPT has been trained on a diverse range of Internet text. However, it does not know any information about specific documents or sources in its training set .

Is AI like Wikipedia in digital format?

Wikipedia is an online encyclopaedia with entries that are updated and edited by its user community. It is structured and organized in a specific manner, with entries for different topics, references, and citations for verification purposes. On the other hand, AI like ChatGPT is not structured in the same way. It doesn't have access to specific databases or encyclopaedias, including Wikipedia, and can't pull information directly from them. Wikipedia's content is written by human users around the world, who bring their understanding and interpretation to the articles they create or edit. In contrast, ChatGPT generates responses based on patterns and information it learned during its training on a broad dataset of Internet text.

How long will AI last or evolve for the lab? We all know that today's marvelous technology becomes obsolete quickly. How do we keep up?

AI is a rapidly developing field, with advancements and innovations happening frequently. New algorithms and architectures, more powerful computational resources, and larger and more diverse training datasets continually push the boundaries of what AI can do.

This is a big investment for some. By investing in a product like this freezer, how much money can a company save per year by integrating this equipment into their lab? How long will it take to realize the return?

Here are some things to consider:

Energy efficiency: If the new freezer is more energy-efficient than the existing ones, you'll save on energy costs. You'd need to compare the energy consumption of the old vs. new freezer and calculate the savings based on your local energy costs.

Maintenance costs: Newer equipment might require less frequent maintenance or be less likely to break down, leading to savings on repair and maintenance costs.

Labor costs: If the new freezer has features that make lab workers more efficient, such as better organization or faster cooling times, it could lead to labor cost savings.

Longevity: High-quality equipment might last longer before needing to be replaced, adding to long-term savings.

Should it also help us to anticipate and prevent failure?

AI-driven potential failure prevention: In addition to the dual redundant refrigeration system and cloud monitoring capability, the North Sciences/Traceable® TEC2 Ultra-Low Temperature Freezer leverages the power of AI to detect potential failure scenarios and prevent them from occurring. Through continuous data analysis and machine learning algorithms, the AI system identifies patterns, trends, and anomalies in the freezer's operational data. By proactively recognizing signs of potential malfunctions or deviations from normal behavior, the AI system can take corrective actions or alert you to the situation. This early warning system acts as a safeguard, allowing you to address any emerging issues before they escalate and safeguarding your samples against unforeseen failures.

With the North Sciences/Traceable® TEC2 Ultra-Low Temperature Freezer, you can rest easy knowing that your samples are protected by a state-of-the-art system designed to minimize the risk of loss or compromise. The dual redundant refrigeration system ensures continuous operation, while cloud monitoring and AI-driven potential failure prevention provide real-time insights and proactive alerts. Our goal is to ensure your valuable samples are preserved with the highest level of security, allowing you to focus on your research and other critical tasks without worrying about sample integrity.

With all the data that will be created, how do you think this data will be shared and will it be shared in a valuable way? Are there barriers to overcome in that area?

With the increasing amount of data being generated, the sharing of data has become a significant challenge. While it's difficult to predict exactly how data will be shared in the future, there are several trends and considerations that can shed light on this topic, for example data sharing pPlatforms: We can expect the emergence of specialized data sharing platforms that facilitate the exchange of data between individuals, organizations, and even across industries. These platforms may offer standardized formats, protocols, and tools to ensure seamless data sharing.

Do you have concerns about the integrity of the data worldwide that AI would be accessing knowing that some tests can be flawed and resulting in flawed data?

Yes, concerns about the integrity of data worldwide are valid, particularly when it comes to AI accessing and utilizing that data. Flawed data can significantly impact the performance and outcomes of AI systems, leading to biased or inaccurate results. There are several reasons why data integrity can be a concern:

Biased data: Biases can be present in data due to various factors, including sampling methods, data collection processes, and societal biases. If AI systems are trained on biased data, they may perpetuate or amplify those biases, leading to unfair or discriminatory outcomes.

Incomplete or inaccurate data: Data may be incomplete or contain inaccuracies due to errors in data collection, data entry, or other factors. This can result in misleading or incorrect conclusions when AI algorithms are trained on such data.

Data manipulation: Intentional manipulation of data can occur, where individuals or organizations modify data to serve their interests or agendas. This can undermine the reliability and integrity of the data used by AI systems, leading to skewed results.

Data privacy breaches: Unauthorized access or data breaches can compromise the integrity of data. If sensitive or private information is accessed or altered, it can have severe consequences for individuals and the validity of AI systems trained on that data.

If you were to boil down the main pain point of a person who buys your equipment, what would this be? How does AI solve this pain?

Simply, AI brings intelligence, automation, and predictive capabilities to the North Sciences/Traceable® TEC2 ULT Freezer, addressing the pain points associated with preserving valuable samples. By leveraging AI, you can have enhanced confidence in the reliability, efficiency, and security of your samples stored in the freezer.

What tools are available to the public to create AI for our own specific lab needs?

A few tools are:

TensorFlow: TensorFlow is an open-source machine learning framework developed by Google. It provides a comprehensive ecosystem for building and deploying AI models. TensorFlow offers libraries, APIs, and tools for tasks such as data preprocessing, model development, and deployment.

PyTorch: PyTorch is another popular open-source deep learning framework that provides a dynamic computational graph. It offers a user-friendly interface for building and training AI models, making it widely used in research and development. PyTorch provides extensive support for tasks like computer vision and natural language processing.

Keras: Keras is a high-level neural networks API that can run on top of TensorFlow, Theano, or CNTK. It simplifies the process of building and training deep learning models by providing a user-friendly and intuitive interface. Keras is known for its simplicity and flexibility, making it suitable for beginners and advanced users alike.

Scikit-learn: Scikit-learn is a popular Python library for machine learning. It provides a wide range of algorithms and tools for tasks like classification, regression, clustering, and dimensionality reduction. Scikit-learn offers a user-friendly interface and comprehensive documentation, making it an excellent choice for building AI models in various domains.

AutoML tools: Automated Machine Learning (AutoML) tools aim to simplify the process of building AI models by automating tasks like feature engineering, model selection, and hyperparameter tuning. Popular AutoML tools include Google Cloud AutoML,, and Auto-Keras. These tools are designed to streamline the AI development process, even for users without extensive machine learning expertise.

Jupyter Notebook: Jupyter Notebook provides an interactive environment for developing and sharing code. It lets you combine code, visualizations, and explanatory text in a single document. Jupyter Notebook is widely used for AI development, experimentation, and sharing of research.

I would like to use AI to compile data similar to your literature review example. How does one get started building an AI bot? What tools are available to the public that we can use to build this AI ourselves?

Remember, building an AI bot requires a solid understanding of machine learning concepts and programming skills. Some tools available to the public for building AI bots:

Python: Python is a widely used programming language for AI development due to its extensive libraries and frameworks. It provides flexibility, ease of use, and a large developer community. Python libraries like TensorFlow, PyTorch, and scikit-learn offer rich functionality for building AI models.

Natural Language Processing (NLP) Libraries: NLP libraries such as NLTK (Natural Language Toolkit), spaCy, and Gensim provide functionalities for text processing, language modeling, and other NLP-related tasks. These libraries can be helpful for tasks like text extraction, summarization, and sentiment analysis.

Chatbot Frameworks: Frameworks like Rasa, ChatterBot, and DialogFlow specialize in building conversational AI agents or chatbots. These frameworks provide prebuilt components and tools for natural language understanding, dialogue management, and response generation.

Cloud AI Services: Cloud providers like Google Cloud, Amazon Web Services (AWS), and Microsoft Azure offer AI services, including pretrained models, APIs, and infrastructure for building AI applications. These services can help simplify the development and deployment process.

OpenAI GPT-3 API: OpenAI provides an API for accessing their language model GPT-3. You can leverage this API to build AI bots that generate human-like text or perform language-based tasks.

Can you please share a few practical use-case of using AI in a lab environment?

Certainly. Here are a few practical use-cases of using AI in a lab environment:

Image analysis and object recognition: AI can be used to analyze images captured from microscopes or other imaging devices. It can automatically detect and classify cells, organisms, or structures of interest, saving time and reducing human error. For example, AI algorithms can identify cancer cells in histopathology slides or analyze bacterial growth patterns in petri dishes.

Drug discovery and development: AI can expedite the drug discovery process by analyzing vast amounts of biological and chemical data. It can predict the properties and interactions of molecules, identify potential drug targets, and optimize lead compounds. AI models can also aid in simulating drug effects, predicting toxicity, or screening drug candidates, thereby accelerating the development of new therapies.

Laboratory automation and robotics: AI can be integrated with robotic systems to automate repetitive and labor-intensive tasks in the lab. This includes sample handling, pipetting, plate labelling, and data entry. By using AI, laboratories can increase efficiency, minimize human errors, and improve throughput in processes such as high-throughput screening or sample preparation.

Quality control and anomaly detection: AI can monitor and analyze sensor data from lab equipment to detect anomalies or deviations from expected patterns. For instance, it can identify irregularities in temperature, pressure, or other variables, indicating potential equipment malfunctions or sample degradation. This allows for proactive maintenance, ensuring the accuracy and reliability of experimental results.

Predictive analytics for experimental outcomes: AI models can be trained on historical experimental data to predict outcomes or optimize experimental conditions. By leveraging patterns and correlations in data, AI algorithms can suggest optimal parameters, identify potential experimental pitfalls, or recommend alternative approaches. This enables scientists to make informed decisions, reduce trial and error, and optimize resources.

Literature and data mining: AI can assist in searching and analyzing vast scientific literature or databases, helping researchers identify relevant studies, extract key information, and summarize findings. AI-powered tools can also provide personalized recommendations, enabling scientists to stay updated with the latest research and make evidence-based decisions.

These are just a few examples of how AI can be applied in a lab environment.

We are a commercial service lab and our workflow is from receiving samples to sharing results; I am looking for specific examples of using AI in the whole laboratory workflow.

Here are some specific examples of using AI throughout the workflow in a commercial service lab:

Sample identification and tracking: AI can assist in automating sample identification and tracking processes. By integrating AI with barcode or image recognition systems, the lab can accurately identify and track samples throughout the workflow, reducing manual errors and ensuring sample integrity.

Intelligent sample prioritization: AI algorithms can analyze sample characteristics, urgency, and specific requirements to prioritize the order of processing. This helps optimize resource allocation, minimize turnaround times, and ensure timely delivery of results for critical or time-sensitive samples.

Automated data extraction and entry: AI can automate the extraction of relevant information from sample submission forms or laboratory reports. Natural Language Processing (NLP) techniques can be employed to extract data from unstructured text and enter it into the laboratory information management system (LIMS), reducing manual data entry efforts and improving efficiency.

Data analysis and interpretation: AI can be used to analyze complex data generated by laboratory instruments, such as high-throughput sequencing, mass spectrometry, or chromatography. AI algorithms can identify patterns, correlations, or anomalies in the data, assisting in data interpretation, result validation, and quality control.

Predictive analytics for quality control: By training AI models on historical quality control data, it becomes possible to predict potential issues or deviations in quality control measures. AI can identify patterns indicating quality variations, instrument drift, or sample contamination, enabling proactive interventions and ensuring high-quality results.

Customized reporting and insights: AI can generate customized reports tailored to specific client requirements or regulatory standards. AI algorithms can automatically analyze data, generate visualizations, and generate insightful summaries, providing clients with comprehensive and easily interpretable results.

Customer support and query resolution: AI-powered chatbots or virtual assistants can handle customer inquiries, provide real-time status updates on sample processing, and answer common queries. This reduces the need for manual intervention, enhances customer satisfaction, and streamlines communication channels.

Resource optimization and workflow efficiency: AI can optimize the allocation of laboratory resources, including personnel, equipment, and reagents. By analyzing historical data and workflow patterns, AI algorithms can provide recommendations for workload distribution, equipment utilization, and resource planning, ensuring efficient laboratory operations.

What kind of benefits does AI provide in the field of quality control?

AI offers several benefits in the field of quality control:

Enhanced accuracy: AI algorithms can analyze large volumes of data with precision and accuracy, minimizing human error. By automating quality control processes, AI can detect defects or anomalies in products or processes that may be difficult to identify through manual inspection alone.

Increased efficiency: AI can significantly speed up quality control processes by automating tasks that would otherwise require substantial human effort and time. This allows for faster and more efficient inspection, testing, and analysis, leading to improved overall productivity.

Early detection of quality issues: AI algorithms can detect subtle patterns or deviations in data that may indicate quality issues or potential failures. By continuously monitoring data from sensors, instruments, or production lines, AI can provide early warnings and trigger corrective actions, minimizing the impact of quality issues and reducing waste or rework.

Predictive maintenance: AI can analyze historical data from equipment sensors and maintenance records to predict when machinery or instruments may fail or require maintenance. By identifying patterns and anomalies, AI can recommend timely maintenance or servicing, reducing downtime and ensuring consistent quality control.

Process optimization: AI can identify optimization opportunities in quality control processes. By analyzing data from various stages of production or inspection, AI algorithms can identify bottlenecks, inefficiencies, or areas for improvement. This enables data-driven decision-making and process optimization to enhance overall quality control effectiveness.

Data-driven insights: AI can analyze large datasets from quality control processes, identifying trends, correlations, and root causes of quality issues. By providing actionable insights and visualizations, AI empowers quality control professionals to make informed decisions, improve processes, and drive continuous improvement initiatives.

Reduced costs: Through automation and early detection of quality issues, AI can help reduce costs associated with rework, scrap, or product recalls. By minimizing defects and improving product quality, AI contributes to cost savings and improved profitability.

Regulatory compliance: AI can assist in ensuring compliance with quality standards and regulatory requirements. By analyzing data and identifying deviations or noncompliant patterns, AI helps maintain adherence to quality regulations, reducing the risk of compliance issues or penalties.

Do you have any advice on where to start for a chemist looking to get into AI coding and lab interactions?

Gain familiarity with AI concepts: Start by familiarizing yourself with fundamental AI concepts such as machine learning, deep learning, and neural networks. Understand the basics of how AI models are built, trained, and deployed.

Learn Python programming: Python is widely used in the AI and data science communities. Familiarize yourself with Python programming language, as it is commonly used for AI coding. Start with learning the basics of Python syntax, data structures, and control flow.

Explore AI libraries and frameworks: Dive into popular AI libraries and frameworks such as TensorFlow, PyTorch, and scikit-learn. These frameworks provide a wide range of tools and resources for building AI models, and they have extensive documentation and online tutorials to get you started.

Take online courses: Online platforms like Coursera, Udacity, and edX offer a variety of AI-related courses. Look for courses specifically designed for chemists or those interested in applying AI in chemistry-related fields. These courses will provide you with structured learning paths, practical exercises, and projects to gain hands-on experience.

Leverage chemical informatics tools: Explore chemical informatics tools and libraries that integrate AI capabilities, such as RDKit, Cheminformatics Toolkit, or Chemoinformatics Workflows. These tools are specifically designed for chemists and enable AI-driven analysis and modelling of chemical data.

Collaborate with data scientists or AI experts: Engage with data scientists or AI experts within your organization or research community. Collaborate with them on projects or seek guidance in applying AI techniques to your specific research or lab tasks. Their expertise can provide valuable insights and accelerate your learning process.

Start with small projects: Begin by applying AI to small-scale projects within your lab or research area. This could involve analyzing existing data, building simple AI models, or integrating AI into specific lab processes. Starting with manageable projects allows you to gain practical experience and gradually expand your AI capabilities.

Stay updated with research and industry trends: Stay up to date with the latest advancements in AI applications within the field of chemistry. Follow relevant research papers, attend conferences, and join online communities to stay informed about emerging AI techniques and their applications in the lab environment.

Remember, learning AI coding and its application in lab interactions is an ongoing process.