AI ethics scores in automated orchestration decision-making
One example method includes receiving an orchestration automation request for an asset, identifying an ethics rule that applies to the asset, comparing the ethics rule to asset values contained in an AI ethics datastore, based on the comparing, identifying a list of assets that conform to the ethics rule, and when the asset appears in the list of assets that conform to the ethics rule, automatically placing the asset at an entity of a computing infrastructure.
1. A method, comprising:
receiving an orchestration automation request for an asset;
identifying an ethics rule that applies to the asset;
comparing the ethics rule to asset values contained in an AI ethics datastore, wherein the ethics rule relates to AI ethics pillars, and the AI ethics pillars comprise (1) accountability, (2) value alignment, (3) explainability, (4) fairness, and (5) user data rights;
based on the comparing, identifying a list of assets that conform to the ethics rule;
using a historian, reviewing prior ethics-based decisions concerning placement and use of the asset;
in response to the reviewing of the prior ethics-based decisions, triggering a notification that a risk assessment concerning the asset should be performed by a prospective user of the asset prior to use of the asset by the prospective user;
based on the list of assets that conform to the ethics rule and the reviewing of the prior ethics-based decisions, automatically placing the assets at an entity of a computing infrastructure; and
measuring a score for each of the AI ethics pillars by context of the asset, a machine, a creator of the asset, and a peer.
2. The method as recited in claim 1 , wherein the triggering is performed when historical placement decisions involving the asset have been reevaluated and a determination made that the asset no longer conforms with the ethics rule.
3. The method as recited in claim 1 , wherein when the asset does not appear in the list of assets that conform to the ethics rules, the list of assets and the asset values are referred to a secondary decision making process.
4. The method as recited in claim 3 , wherein the secondary decision making process is performed based in part on input concerning placement of the asset.
5. The method as recited in claim 1 , wherein placing the asset at the entity comprises placing a computing workload at a particular computing system.
6. The method as recited in claim 1 , wherein the ethics rule is defined and specified by an organization that orchestrates placement of the asset.
7. The method as recited in claim 1 , wherein the AI ethics datastore maintains a list of assets of an organization with which the AI ethics datastore is associated.
8. The method as recited in claim 1 , wherein the asset is associated with an asset context defined by a user.
9. The method as recited in claim 1 , wherein the ethics rule comprises a specified AI ethics tally that indicates a minimum ethical standard to which the asset must conform.
10. A non-transitory storage medium having stored therein instructions that are executable by one or more hardware processors to perform operations comprising:
receiving an orchestration automation request for an asset;
identifying an ethics rule that applies to the asset;
comparing the ethics rule to asset values contained in an AI ethics datastore, wherein the ethics rule relates to AI ethics pillars, and the AI ethics pillars comprise (1) accountability, (2) value alignment, (3) explainability, (4) fairness, and (5) user data rights;
based on the comparing, identifying a list of assets that conform to the ethics rule;
using a historian, reviewing prior ethics-based decisions concerning placement and use of the asset;
in response to the reviewing of the prior ethics-based decisions, triggering a notification that a risk assessment concerning the asset should be performed by a prospective user of the asset prior to use of the asset by the prospective user;
based on the list of assets that conform to the ethics rule and the reviewing of the prior ethics-based decisions, automatically placing the assets at an entity of a computing infrastructure; and
measuring a score for each of the AI ethics pillars by context of the asset, a machine, a creator of the asset, and a peer.
11. The non-transitory storage medium as recited in claim 10 , wherein the triggering is performed when historical placement decisions involving the asset have been reevaluated and a determination made that the asset no longer conforms with the ethics rule.
12. The non-transitory storage medium as recited in claim 10 , wherein when the asset does not appear in the list of assets that conform to the ethics rules, the list of assets and the asset values are referred to a secondary decision making process.
13. The non-transitory storage medium as recited in claim 12 , wherein the secondary decision making process is performed based in part on input concerning placement of the asset.
14. The non-transitory storage medium as recited in claim 10 , wherein placing the asset at the entity comprises placing a computing workload at a particular computing system.
15. The non-transitory storage medium as recited in claim 10 , wherein the ethics rule is defined and specified by an organization that orchestrates placement of the asset.
16. The non-transitory storage medium as recited in claim 10 , wherein the AI ethics datastore maintains a list of assets of an organization with which the AI ethics datastore is associated.
17. The non-transitory storage medium as recited in claim 10 , wherein the asset is associated with an asset context defined by a user.
18. The non-transitory storage medium as recited in claim 10 , wherein the ethics rule comprises a specified AI ethics tally that indicates a minimum ethical standard to which the asset must conform.