A Weekly Upskill for Yourself, Your Data, Your Team

September 24, 2026

Issue 130

Implementing AI isn't just about your tool stack

Successfully implementation AI across your organization (aka getting the outcome you set out for) requires more than powerful tools.

Most enterprise AI projects fail because they are built in isolation.

• Someone builds a cool AI project without considering the company setup it will connect to.
• They try to connect data to a third-party software before they talk to legal (hello, privacy policies).
• A tool gets built for one person and works beautifully, until someone asks about rolling it out to the enterprise and realizes the cost and complexity never penciled out at scale.
• An AI feature gets designed and delivered, and the developers say integration is eighteen months out.
• Or, they spend six months accidentally building Sticker AI, thinking they are building ROI AI, not realizing the difference, and end up with neither.

Countless AI features get demo-ed, applauded, and forgotten because building goes first and integration follows as a loose afterthought.

If you're in the mist of building a data and AI project (or before you start the next one) consider:

• What is the intended outcome for this project? ROI-generating, time-saving, creating buy-in from stakeholders?
• Which existing company systems and tool should the project connect to?
• Who else needs to be consulted? Legal, dev team?

And this one often gets ignored:

• How are you actively creating a company culture and policies around AI?

Curious what successful AI implementation looks like in action? See how DMSi used Sticker AI to build company-wide buy and then built one ROI AI project that saved 2,496 working hours in one year.​

Can AI be sued when it gives the wrong answer?

The courts haven't decided but we are sure to see more case law in the next few years, and it will likely revolve around Section 230.

Established in 1996, Section 230 of the Communications Decency Act (“Section 230”) protects platforms that provide user-posted content from being sued or held liable for the content that its users posts.

Section 230 has been used to shield social media companies, such as Meta and X (when it was Twitter) from liability for its algorithms showing certain content, such as terrorists organizations. In 2023, Supreme Court "found that the 'passive' algorithm served as a delivery 'pipe' rather than a 'post,' operating at an arm’s length and 'agnostic' to content, solely transmitting information made by users."

Now, the question becomes:

Does Section 230 shield GenAI as well?

So far, it depends. Crowell summarizes the current court standings like this:

• "Section 230 protects platforms that deliver user content neutrally, but may not protect platforms whose GenAI tools evaluate, transform, or generate new outputs from that content.
• Courts are already drawing the line: using GenAI to moderate content may be protected, but using it to enhance, revise, or proactively recommend content based on user metadata may not be."

And gives the added advice:

• "Platforms integrating GenAI should prioritize content neutrality, separate hosting from generating, and label and document AI outputs now before litigation forces the analysis."

​Keep reading "The Pipe, Not the Posts: How Section 230’s Protections Extend to Generative AI Platforms"​

From the CAN Slack Channel

OpenAI's Ad Tracking

Just like Meta and Google tracking, companies who run ads on OpenAI can now add OpenAI tracking to their website. This means OpenAI can connect your activity on other websites to your account.

​See how it works. ​

New Discovery at the Anthropic Biology Lab

The same month it shared a report containing biological misuse, Anthropic announces its own biology research lab.

"Our focus is on fundamental biology research using Claude: exploring datasets of DNA to identify uncharacterized protein families, generating hypotheses at scale, and testing them through experiments in the lab." - Anthropic

Read about the labs first published finding: Claude discovers a novel enzyme system with CRISPR-like repeats​

DoorDash's in-house AI data-analysis

DoorDash built an in-house AI analysts because they "found that building and tuning all parts of the pipeline — prompting, ranking and retrieval, and domain-specific skills — led to significant performance improvements over standard agents with simple data connectors."

​As Ron Potok points out on LinkedIn, there is something to be gleamed from DoorDash's strategy to build your own AI infrastructure, even if you aren't 10,000 users deep:

This is exactly what we help our clients build: the harness, the layers underneath and the internal decision making ability to know when to build versus rent. Not sure how to start building your data and AI infrastructure? Get in touch today.

Analytics After Hours @ The Tipsy Pig

September 30, 4-7:00 PM

A message from Nebraska Data Analytics Forum

​RSVP on Meetup!​