22 Aug 2023
AI Code of Ethics & Acceptable Use Policy for Teammate AI
Emphasize Uniqueness & Originality
AI is used at Teammate AI to free up humans to be more unique, original, and creative. We do not outsource or replace the characteristics that make us unique to AI.
AI Works for Humans
AI is used as a tool for humans. Humans are not used as a tool for AI. We keep the humans at the forefront of our intention with the purpose of freeing them up.
Supervised to Unsupervised
Unsupervised AI only becomes client-facing if we have confidence that the unsupervised AI does a task markedly better than a human would.
Pure Transparency
Every Teammate AI-built system must clearly show how and why it makes decisions. This could be through easy-to-understand logs or by letting us look into the workings of our neural networks. Being clear about how decisions are made helps us improve fairness, data quality, and performance.
AI Do’s and Don’ts
Do:
- Experiment with AI tools to enhance your workflow and your teamwork.
- Investigate Privacy Policies of any AI tools you might use.
- Feel comfortable using Microsoft products and AI tools.
- Feel comfortable using Teammate AI-built AIs at your discretion with confidence regarding data security.
- Ask the leadership team for their opinion on new AI tools that can enhance our workflows or teamwork.
- Add extra logging & transparency frameworks into our systems so that we can continuously monitor and improve.
Don’t:
- Use an AI to replace yourself in a client interaction.
- Use any AI service which grants a royalty-free license on any data you upload or submit.
- Upload any sensitive information to an AI. For example, if you would not post the information on the public Teammate AI website, do not upload it to an AI that is not owned by Teammate AI.
- Grant any global read-access on your local machine to an AI software. Tools such as Rewind.AI are forbidden.
- Train or fine-tune any custom AI without prior written consent from the leadership team.
- Expect that shipping a “black-box” solution is acceptable. We cannot improve what we cannot measure.
22 Aug 2023
AI Code of Ethics & Acceptable Use Policy for Teammate AI
Emphasize Uniqueness & Originality
AI is used at Teammate AI to free up humans to be more unique, original, and creative. We do not outsource or replace the characteristics that make us unique to AI.
AI Works for Humans
AI is used as a tool for humans. Humans are not used as a tool for AI. We keep the humans at the forefront of our intention with the purpose of freeing them up.
Supervised to Unsupervised
Unsupervised AI only becomes client-facing if we have confidence that the unsupervised AI does a task markedly better than a human would.
Pure Transparency
Every Teammate AI-built system must clearly show how and why it makes decisions. This could be through easy-to-understand logs or by letting us look into the workings of our neural networks. Being clear about how decisions are made helps us improve fairness, data quality, and performance.
AI Do’s and Don’ts
Do:
- Experiment with AI tools to enhance your workflow and your teamwork.
- Investigate Privacy Policies of any AI tools you might use.
- Feel comfortable using Microsoft products and AI tools.
- Feel comfortable using Teammate AI-built AIs at your discretion with confidence regarding data security.
- Ask the leadership team for their opinion on new AI tools that can enhance our workflows or teamwork.
- Add extra logging & transparency frameworks into our systems so that we can continuously monitor and improve.
Don’t:
- Use an AI to replace yourself in a client interaction.
- Use any AI service which grants a royalty-free license on any data you upload or submit.
- Upload any sensitive information to an AI. For example, if you would not post the information on the public Teammate AI website, do not upload it to an AI that is not owned by Teammate AI.
- Grant any global read-access on your local machine to an AI software. Tools such as Rewind.AI are forbidden.
- Train or fine-tune any custom AI without prior written consent from the leadership team.
- Expect that shipping a “black-box” solution is acceptable. We cannot improve what we cannot measure.