I read the current version of the European Union Artificial Intelligence Act (AIA) over the weekend.
That might be reason for scorn – or pity – especially as the weekend skies were blue here in Oregon.
Let me share three takeaways from the perspective of an interested observer:
Takeaway 1: like the EU General Data Protection Regulation (GDPR), this will shape the vocabulary, operative definitions, and acceptability of the use of voice and conversational AI across the western world. It’s the Brussels Effect again in action.
As you know, some artificial intelligence usage will be prohibited. Some usage will be defined as high risk. High risk usage will be acceptable only in certain situations and with precautionary processes in place.
If you don’t know the vocabulary and the definitions, you’ll be putting yourself, your business, and your clients at risk.
Because…
Takeaway 2: even if you don’t know the details, you can be sure that your clients’ legal and security officers – and the legal and security officers at their clients – will know the details, word for word.
Their questions and assessment of risk and potential harm will determine whether your deal lives or dies. Perceived risk beats ROI every time.
Especially if you’re involved – even a little bit – with biometric identification, voice authentication, voice or media synthesis, sentiment analysis, or personal attribute inference.
Especially if your work faintly sounds like the uses listed in the Prohibited or High Risk categories.
Which leads to…
Takeaway 3: proactive is better than reactive.
Read the AIA. Yes, I know it’s 107 pages. Get a full carafe of coffee and do it.
Review it with your attorney. Carefully.
Document why and how you’re on the good side of the AIA – so, when you're challenged (and you will be), you can speak clearly and factually when the legal and security heavyweights step into the room.
Especially if you’re in one of the categories listed above.
And contact Oita Coleman at the Open Voice Network, and prepare to enroll with your team in the TrustMark Initiative voice/conversational AI ethics course. And to work your way through the ethical use self-assessment that you’ll soon find on our TrustMark Initiative page.
The TrustMark Initiative: ethical principles translated into action. Risk mitigation for you and your team.
The EU AIA is yet another inflection point for the world of voice and conversational AI. The rules of engagement just changed.
