AI-agents should have ID-wallets

Why?  

Because they are of limited use unless they can be trusted and be interoperable with your own organisational and personal ID-wallet and other AI-agents. Your customers, suppliers, the public sectors and your staff will have ID-wallets and thus be able to interact with them as instructed.  

It should also be underlined that even if AI agents can locate data and further automation on a grand scale they cannot enter into legally binding agreements. The ultimate responsibility for the AI-agent's transactions lies with natural and legal persons.  This will happen by interchanging verifiable data and using ID-building wallets also for signing transactions. 

To me it appears clear that: 

1. At the deployment stage you can see the AI-agent as an employment seeker. It should thus have a set of verifiable credentials like human applicants – and have the same interoperable ID-building wallets for interchanging data (without technical integration).
2. You expect it to be able to present an identification credential – verifiable with the organisation offering the AI-agent as a service (for signing the “lease” contract) or as a purchase (for “warranty” and security)
3. When you use the AI-agent to find data only inside the organisation the security aspect is not as paramount as otherwise – but as credentials collected from outside and from staff and to be sent outside pass through your organisation wallet it appears clear that the AI-agent uses its wallet to interact when collecting data for this part
4. When you send your AI-agent out for specific tasks you need to give it the needed credentials just like your staff need to have – starting from identification and power to act credentials for  a specific - often time-limited mission. And be able to trust that the credentials are used in a way that you can trust and only for the intended purpose.
5. The parties (and their AI-agents) your AI-agents then meets,  expect it to present the verifiable identification, your power-to-act credential, other needed data and probably also security related classification credentials – from wallet to wallet
6. When you – or your AI-agent -  are approached by other parties (staff) or their AI-agents you of course expect them to present credentials that you now finally can be easily verify in a unified wallet-to-wallet-to-wallet automated way 
7. As collections of data – more or less strongly verified (all the way to open data and real time behavioral data) – is needed by you, your staff and customers for business and life events it appears clear that the task of finding it and getting it to your wallet and passing it on in a way that you control demands smart both the AI-agents and the trust that only ID-wallets can bring
8. This is needed for agent2agent https://www.linkedin.com/posts/bo-harald-4768b51_ai-agents-mcp-activity-7316733129812013057-MhKc?utm_source=share&utm_medium=member_desktop&rcm=ACoAAABPj1oB9_D7YNYACmHvY9HioUqpuULqZCo

Maybe this is all too self-evident by now? Please feel free to improve – or show another direction to the data driven economy.