In traditional review, bias might mean one reviewer tags Executive emails as “more relevant” by default (authority bias) or overlooks documents written in non-standard English (linguistic bias). In CAL, those coding decisions become training data, directly steering the machine’s predictions.

In GenAI-assisted review, the risks increase. The model can take these inputs, generate skewed summaries or coding rationales, and reinforce that across thousands (or millions) of documents. More simply put, biased reviewers create biased training data which creates biased AI outputs.

Consider a typical scenario where a review team is analyzing internal communications during a corporate investigation. If reviewers unconsciously treat technical staff emails as less relevant than executive communications, the CAL model learns to deprioritize those types of documents. When a GenAI tool later generates summaries or rationales, it mirrors that bias, highlighting executive language and downplaying other perspectives. Over time, that becomes a systemic pattern, not a one-off error.

Bias can also emerge when reviewers make cultural or linguistic assumptions. A phrase that sounds sarcastic in one region may be neutral in another. If those cues aren’t calibrated, the AI learns misaligned signals. This is how subtle human inconsistency becomes amplified into machine-level bias.

So, how do forward-looking legal teams embrace new possibilities without abandoning their foundational need for reliability? They do it by pairing bold ideas with strategic, disciplined frameworks, which is exactly what we have done at TCDI with our Tech Lab’s approach to innovation and evaluating emerging technologies.

As our clients and teams identify needs, we have to remain committed to developing and testing new legal technology and workflows to ensure we deliver innovative, secure, and effective solutions. The structure of TCDI’s Tech Lab has allowed us to do just that.

For the past few years, firms and legal departments have been like excited kids on Christmas morning, unboxing every shiny GenAI feature they could find. They piloted multiple tools and experimented without fully integrating solutions into workflows. That experimentation was important. It helped organizations understand risks and really figure out what AI could (and couldn’t) do.

But now the industry is entering a different phase. The focus is shifting from experimentation to choosing with purpose, implementing with structure, and measuring with intention. 

It’s no surprise that artificial intelligence was front and center. Legal departments are all at different stages of their journey. Some are fully integrating AI into their everyday work, while others remain in pilot phases or watching from the sidelines.

Several attendees shared that their organizations have now received clear direction from leadership: to implement AI to drive measurable efficiency. We heard examples ranging from legal playbooks to supply chain management and M&A due diligence to contract drafting pilots.

Others were more cautious, raising questions about governance, data security, and reliability. Still, everyone agreed on one thing: AI is here to stay, and the conversation has shifted from “Should we use it?” to “How do we use it responsibly and effectively?”

Unsurprisingly, eight of the sessions referenced AI in the title or description, which tells you just how sweeping the topic has become.  Just as families argue over whether stuffing belongs in the turkey (heads up, it should be on the side), the debate around when and how to use AI was as lively as any dinner table could be.   

Among the cornucopia of topics were conversations on how GenAI supports witness profiles, chronologies, and case-building, all with a generous helping of practical tips. We were able to hear expert perspectives and predictions on a mix of AI issues impacting eDiscovery, including:

• The impact of AI on privilege: is there privilege between a non-attornies and AI?
• Are prompts protected work product?
• Should communications with AI be logged?
• Is AI-generated content a business record? If so, who is the custodian?
• What is the discovery and liability impact with the shift to Agentic AI?

And layered across these conversations was a new reality that came through loud and clear. Courts expect companies to understand how their AI tools create, store, and manage data. There is no holiday-week grace period on this one. If AI touches your workflow, you will be expected to account for it during litigation.