From Lab Automation to AI Scientists 

What AI orchestration of laboratory instruments means for the future of lab automation 

Key Takeaways 

• AI is moving from analysis to orchestration: AI/ML implementation is extending beyond data analysis and workflow support into orchestration of laboratory instruments and automated systems. Anthropic’s Model Hardware Standard (MHS), a research preview from August 2026, lets AI agents discover and operate lab equipment with safety controls enforced at the device level. 

• AI is already running experiment loops: In February 2026, OpenAI and Ginkgo Bioworks reported that GPT-5 proposed and analyzed rounds of experiments across more than 36,000 reactions in an automated lab, with OpenAI citing a 40% reduction in protein production costs. 

• Near-term adoption stays human-supervised: Most laboratory organizations are expected to maintain human oversight, with near-term AI adoption centered on human-supervised and modular automation rather than fully autonomous labs. 

• Researchers are growing more cautious: Wiley’s 2025-2026 study found concerns about inaccuracies, hallucinations, privacy, and security rose alongside AI usage. In an MIT quantum calibration case study, expert human intervention was often needed when data quality deteriorated. 

• Security and regulation are still taking shape: New attack vectors targeting AI agents are under evaluation, the EU AI Act exempts AI developed exclusively for scientific R&D, South Korea requires AI usage disclosure, and the US has no comprehensive federal AI law. 

Artificial intelligence is playing an increasingly important role in laboratory automation, supported by growing collaboration between laboratory technology suppliers and AI developers. However, automation adoption rates and frameworks vary broadly. Developments over the past few years suggest AI/ML implementation beyond data analysis and workflow support into orchestration of laboratory instruments and automated systems.  

One example is Anthropic’s Model Hardware Standard (MHS), introduced as a research preview in August 2026. MHS is a collection of device drivers designed to allow AI agents to discover and operate laboratory equipment while enforcing safety controls at the device level. Early industry participants include Tecan, which plans to support MHS in its Fluent liquid handling platform Automata, which is integrating the standard into its LINQ automation software. Danaher has also indicated interest in exploring the technology for future smart instruments. 

AI is increasingly being applied within automated experimentation workflows. In February 2026, OpenAI and Ginkgo Bioworks reported that GPT-5 was used to propose and analyze multiple rounds of experiments in Ginkgo’s automated laboratory, involving more than 36,000 cell-free protein synthesis reactions. According to a press release from OpenAI, the project contributed to “a 40% reduction in protein production costs.” The AI system designed and evaluated experimental iterations, while robotic systems executed the laboratory work. Such arrangements are increasingly described based on the level of autonomy infused into labs, such as self-driving labs where software and automation systems operate within an iterative experimental feedback loop. 

Despite these advances, near-term AI adoption in lab automation is expected to remain centered on human-supervised and modular automation workflows. Although AI adoption continues to increase across scientific research, researchers are becoming more measured in their expectations of the technology. Wiley’s 2025-2026 study ExplanAItions: The Evolution of AI in Research found that concerns regarding inaccuracies, hallucinations, privacy, and security increased alongside AI usage, suggesting that practical experience is fostering a more cautious and nuanced view of AI’s capabilities. 

The distinction between human-in-the-loop and lab-in-the-loop systems is becoming increasingly important. In a lab-in-the-loop environment, AI systems participate in a closed loop of experimental design, execution, analysis, and optimization, with results continuously informing future experiments. In a human-in-the-loop environment, scientists retain responsibility for reviewing, approving, and modifying AI-generated decisions and recommendations.  

While fully autonomous laboratories continue to attract attention, most laboratory organizations are expected to adopt approaches that maintain human oversight. Recent research reinforces this view; a 2025 review published in Science Robotics identified reproducibility, standardization, governance, ethics, and human oversight as continuing challenges for autonomous laboratories. Similarly, NIST’s AI Risk Management Framework emphasizes the importance of clearly defining human responsibilities when overseeing AI-enabled systems. 

Case studies continue to demonstrate both the potential and limitations of current AI technologies. In 2026, OpenAI highlighted work by graduate students at MIT’s Engineering Quantum Systems group, in which an AI agent was connected to laboratory control software to help calibrate quantum systems. While the agent successfully automated portions of the workflow, researchers reported that expert human intervention was often required when data quality deteriorated or experimental conditions became ambiguous. The study underscored a broader industry trend: AI can accelerate laboratory processes, but experienced scientists remain essential for addressing complex or uncertain situations. 

As AI becomes more tightly integrated with laboratory software and instrumentation, cybersecurity considerations are receiving greater attention. Researchers have begun evaluating new attack vectors that could target AI agents through manipulated tool descriptions, software interfaces, or corrupted training data. Although these risks remain an active area of research, they highlight the importance of governance, validation, and security controls as laboratories expand the use of connected AI systems. 

The regulatory environment surrounding AI is also evolving. In Europe, the AI Act exempts AI systems developed exclusively for scientific research and development, while implementation timelines for certain high-risk AI obligations have been extended under the EU’s Digital Omnibus initiative. In South Korea, the AI Basic Act requires disclosure of AI usage and transparency regarding high-impact AI systems. In the United States, no comprehensive federal AI law has been enacted, although various proposals continue to be debated while state-level legislation expands. 

Overall, the market is moving toward greater integration of laboratory software, AI/ML, and automated hardware. However, the latest evidence suggests that the near-term future is more likely to be characterized by human-supervised and lab-in-the-loop automation than by fully autonomous laboratories. Reliability, governance, security, and regulatory considerations continue to favor models in which AI augments scientific decision-making rather than replaces it. 

SDi’s newly published Laboratory Automation Markets 2026 report examines these developments within the broader automation market, providing forecasts from 2025 to 2030 by technology, end market, and region, along with vendor market shares and supplier participation matrices across eight automation categories: 

• Sample Management 

• Liquid Handling 

• Informatics & Automation Software 

• Microplate Readers 

• Autosamplers, Imaging Stages, and Biosensors 

• LIMS, ELN, and SDMS 

• Multiplex & High-Throughput ELISA 

• Robotics 

The report’s supplementary materials also outline laboratory automation maturity levels with reference to each technology category, providing a simple framework for assessing both current adoption and future automation opportunities. 

Read the full Laboratory Automation Markets 2026 report >>

Recommended Reading 

• South Korea: More Autonomous Labs Planned (IBO): A national consortium plans AI-powered, closed-loop labs running around the clock. 

• Service: An Intervention When the Hum of Lab Instrumentation Falls Silent (SDi): The role of structured service in modern labs integrating complex systems. 

• Beyond the Bench: Mapping the Life Science Customer Experience (BioInformatics): Survey-based insights that include AI adoption. 

Frequently Asked Questions