@cpownall Not what I originally meant, but your scenario is plausible, too. If in algorithmic trading you have AIs buying and selling stocks against each other, why can’t we expect AIs to debate against each other to achieve goal X (where X can be generating confusion, convincing voters, )

If the assessment and labelling would be extended beyond high-risk #AI applications, a lot of vendors claiming to adopt AI when in fact they don’t would have a hard time in EU. https://t.co/J3EXpLfCAf

/20 “@EU_Commission considers at this stage that an objective, prior conformity assessment would be necessary to verify and ensure that certain of the above mentioned mandatory requirements applicable to high-risk #AI applications are complied with”

Mortgage page of your bank website: “Attention: your chance to have a house is being decided by an #AI. If you don’t know what that means or you have concerns about AI, talk to the chatbot AI in the next screen.” https://t.co/2kN7VEg1pc

/19 “citizens should be clearly informed when they are interacting with an #AI system and not a human being. Whilst EU data protection legislation already contains certain rules of
this kind, additional requirements may be called for to achieve the abovementioned objectives.”

/18 Even just the first set of requirements (training data), detailed in the @EU_Commission whitepaper, makes exceptionally hard to compete for #AI players if they have to procure the datasets by themselves. A pre-vetted, EU-provided dataset might be the only chance for startups.

/17 The @EU_Commission has specific requirements for high-risk #AI in the following aspects:

– training data;
– data and record-keeping;
– information to be provided;
– robustness and accuracy;
– human oversight;

/16 The @EU_Commission considers high-risk #AI that is used for:

– “recruitment processes as well as in situations impacting workers’ rights”

– “remote biometric identification and other intrusive surveillance technologies”

The concept of “enterprise-class support” is the commitment of a vendor to maintain a technology for a long period like 3, 5, 10 years for its business customers. Many vendors deploying #AI in consumer products have nothing like that. https://t.co/iAcGrmnpMF

/15 Another great point from the @EU_Commission on the use of #AI in the supply chain: https://t.co/ePaq88h7Kr