CDP Institute

Copyright issues are scaring off potential users of generative AI as it becomes clear how much training data was taken without permission.  Microsoft has now promised to defend its customers from copyright claims related to its AI “copilots” and to pay any fines if they lose.  Adobe made a similar but more limited promise in June.  Other gen AI developers will probably follow suit.

In addition to using data it shouldn’t, gen AI has problems with making up data that doesn’t exist.  Gleen AI just launched with a promise to reduce that problem by checking AI outputs against an enterprise’s own data, which is presumably accurate.  They have $4.9 million in funding.

Retail media networks and digital out of home advertising (DOOH) are definitely related although they’re often discussed separately.  Vistar Media, a programmatic DOOH vendor, has merged the two with a solution that reaches 25,000 retailer screens using its DOOH-tech.  Buyers can select a variety of retail inventory packages linked to different locations.

Customer engagement platform Braze has introduced an expanded set of data integration features, enabling easier data ingestion, identity management, and live connections with cloud databases including Google BigQuery, Amazon Redshift, Databricks and Snowflake.  We haven’t formally assessed whether Braze qualifies as a CDP but it’s beginning to look they tick most or all of the boxes.

There’s no reason to think that HubSpot is a CDP, but they are certainly a major CRM vendor.  So we should probably mention that they’ve added a bushel of AI tools to generate content, support (or replace) customer service agents, and offer predictive insights.  They also have a public beta of their very own generative AI chatbot, currently with 80,000 total users.

Software subscriptions have accelerated martech growth because they make it easier to buy software than conventional licenses.  But CIOs are less enthusiastic about the subscription model.  This survey from Rimini Street found 99% of CIOs and CTO had concerns about subscription licenses, with the top ones being no return on investment and vendor lock-in.

Meta may be just-short-of-criminally irresponsible, but at least they’re consistent.  Having decided that news coverage isn’t important to their business, even though Meta is an important news source for many people, they’ve decided to remove their Facebook News section in the UK, Germany and France.  They haven’t yet blocked links from news sites, as they did in Canada, but that seems certain to follow.

Meanwhile, for your list of Things There’s No Reason for Meta to Do, you can add an “all-in-one multilingual multimodal AI translation and transcription model” that aims to provide a universal speech and text translator across nearly 100 languages. They haven’t actually released the system, but have posted a one-sentence demo and technical resources for researchers.  It’s not like Meta has better places to spend the money, such as improving content moderation or chasing down fake news.  

Fifty.io identifies advertising audience segments using advanced contextual analysis.  They recently teamed with Scope3 to let clients measure the carbon impact of their campaigns.  All good and noble, but the International Energy Agency estimates global energy-related CO2 emissions were over 36.8 billion metric tons in 2022  while Scope3 estimates that digital advertising creates 7.2 million metric tons (or maybe 13.9 million, although I couldn’t find the source of that figure).  That’s roughly 0.02% of the whole, which, one might argue, isn’t enough to matter.  

One of web publishers’ worst nightmares is that generative AI will answer search questions without sending users to their websites.  With global anti-trust authorities already watching it closely, Google has been very cautious in how it handles this.  Their latest approach in their experimental generative search engine is to embed clickable links to source materials within the answers.  Seems like a move in the right direction.

Feisty regulators are also on Microsoft’s case.  Responding to EU criticisms, they are now unbundling the Teams collaboration platform from Office 365.  The change only applies in Europe, since US regulators continue to be much more lax.

The American Staffing Association reports that workers are increasingly worried about how automation and AI will affect their employment.  Most notably: 47% think automation could easily replace their job, up from 27% in a 2017 survey.  But they’re evenly split on how automation will affect their career: 27% say it will help and 26% say it will be disruptive.