[ai-control] Re: Proposed "AI System Inference" and "AI User Input" categories

[Greg Lindahl <[e-mail]>]

... by both of those definitions, an AI model that examines the text of a
webpage and produces a pornography score output is AI inference.

And you are saying we should have a preference which says "I'm shooting
myself in the foot, never show my webpages in a search engine result."

On Sat, Aug 1, 2026 at 2:21 PM Adam and Joanne Gallagher <
[e-mail]> wrote:

> Apologies for the second email. Here is another definition from IBM:
> https://www.ibm.com/think/topics/ai-inference
>
> "In machine learning, AI inference is the act of using a trained AI model
> to make predictions on new data. Essentially, any instance of an artificial
> intelligence model actually generating outputs or making decisions in a
> real-world AI application constitutes AI inference. In simple terms, AI
> inference entails a trained model applying the patterns it learned from its
> training data to infer the correct output for a given input."
>
> AI inference is generating the output.
>
> Adam
>
> On Sat, Aug 1, 2026 at 11:18 AM Adam and Joanne Gallagher <
> [e-mail]> wrote:
>
>> Happy to hear your definition but this is the definition on the Google
>> document:
>> https://cloud.google.com/discover/what-is-ai-inference#how-does-ai-inference-work
>>
>> "Think of it this way: if training is like teaching an AI a new skill,
>> inference is that AI actually using the skill to do a job. It takes in new
>> data (like a photo or a piece of text) and produces an instant output, like
>> a prediction, generates a photo, or makes a decision. This is where AI
>> delivers business value. For anyone building with AI, understanding how to
>> make inference fast, scalable, and cost-effective is the key to creating
>> successful solutions."
>>
>> Adam
>>
>> On Sat, Aug 1, 2026 at 11:14 AM Greg Lindahl <[e-mail]>
>> wrote:
>>
>>> That's not what AI Inference means.
>>>
>>> On Sat, Aug 1, 2026 at 2:13 PM Adam and Joanne Gallagher <
>>> [e-mail]> wrote:
>>>
>>>> No, AI inference is producing AI recipes riddled with errors that use
>>>> our brand name to compete against us in search in the form of AI Overviews
>>>> and AI Mode outputs. I go into this in detail in my post I linked to. Here
>>>> is the link again:
>>>> https://www.inspiredtaste.net/101454/ai-recipes-harming-trust/
>>>>
>>>> To review the definition of AI inference that is discussed on this
>>>> Google document: https://cloud.google.com/discover/what-is-ai-inference
>>>> . AI inference is the computational process happening behind the
>>>> scenes, and the AI Overview is the final text output of the inference
>>>> process. When someone types a query into Google, the trained model takes
>>>> the query and applies what it learned to generate a unique summary. This
>>>> execution phase of taking the prompt and producing an output is by
>>>> definition AI inference. Bottom line, AI inference is doing the harm
>>>> and we don't have preferences to stop this from happening. As a publisher
>>>> of information this is a major problem and the entire open web currently
>>>> has this problem.
>>>>
>>>> Adam
>>>>
>>>> On Sat, Aug 1, 2026 at 11:00 AM Greg Lindahl <[e-mail]>
>>>> wrote:
>>>>
>>>>> > AI inference is what is proactively being used in search at global
>>>>> scales which is causing grave harm to our business.
>>>>>
>>>>> Are you harmed by AI inference intending to suppress pornography in
>>>>> search results?
>>>>>
>>>>> One great thing about the IETF process is that it brings together many
>>>>> communities -- in this case, both publishers and search engine implementers.
>>>>>
>>>>> This particular topic has come up several times in the past.
>>>>>
>>>>> -- greg
>>>>>
>>>>>
>>>>> On Sat, Aug 1, 2026 at 1:19 PM Adam and Joanne Gallagher <
>>>>> [e-mail]> wrote:
>>>>>
>>>>>> I completely agree with Nate on this. AI inference is what is
>>>>>> proactively being used in search at global scales which is causing grave
>>>>>> harm to our business. We currently don't have any inference preferences to
>>>>>> stop it without removing ourselves from AI outputs entirely. With the lines
>>>>>> between organic index search and AI search being blurred at alarming rates
>>>>>> this would be insane for us to do. It is essentially a hostage situation
>>>>>> right now. Here is a detailed post that describes how AI inference is
>>>>>> misrepresenting our brand and competing against us with our content at
>>>>>> scales never before seen. It is wildly unsustainable. Article link:
>>>>>> https://www.inspiredtaste.net/101454/ai-recipes-harming-trust/
>>>>>>
>>>>>> Our business desperately needs preferences to stop this from
>>>>>> happening as soon as possible. Putting aside the training on our content
>>>>>> that has already happened without our permission, inference is what is
>>>>>> causing (doing) harm to our business in the form of AI misinformation.
>>>>>>
>>>>>> Publishers need to have much more control of inference. It should be
>>>>>> a core mission of this working group to define inference preferences and
>>>>>> vocabulary.
>>>>>>
>>>>>> --
>>>>>> Adam Gallagher
>>>>>> Inspired Taste <https://www.inspiredtaste.net/>* - *Easy Recipes for
>>>>>> Home Cooks
>>>>>>
>>>>>> --
>>>>>> ai-control mailing list -- [e-mail]
>>>>>> To unsubscribe send an email to [e-mail]
>>>>>>
>>>>>
>>>>
>>>> --
>>>> Adam and Joanne Gallagher
>>>> Inspired Taste <https://www.inspiredtaste.net/>* - *Easy Recipes for
>>>> Home Cooks
>>>>
>>>>
>>
>> --
>> Adam and Joanne Gallagher
>> Inspired Taste <https://www.inspiredtaste.net/>* - *Easy Recipes for
>> Home Cooks
>>
>>
>
> --
> Adam and Joanne Gallagher
> Inspired Taste <https://www.inspiredtaste.net/>* - *Easy Recipes for Home
> Cooks
>
>