From personal perspective, I consider that at least one decade must pass for seeing the major AI impacts on everyday life. I consider this trend despite the current advancement of AI techniques & algorithms in many activity domains. For instance, in automotive industry, there are currently no cars (in use) having the latest level of autonomy, namely 5. These cars do not require any human intervention and are guided only remotely using AI methods. In information security, the IDPS devices that perform the tasks automatically currently have high rate of false positives. In any of these cases, it is not about only about innovative design; there are unsolved issues that, in general, require more time. The legal framework represents one of the biggest barriers against the wide adoption of AI-based devices used in the daily life. There must be a synchronization between the legal aspects and available technical means. Smart objects and smart networks are nowadays more present than 10 years ago, even if mainly for research purposes. The AI residing behind the smart objects is not fully validated and enhancements are discovered on periodical basis. It is hard in general for AI techniques to deal with contextual information (even if there were developed some learning strategies that require initial data sets for training). The adoption of devices having more autonomy will be done gradually and perhaps this is the most logical approach. It is first needed to make sure the newer technologies are stabilized before launching the in the production environments. This will also allow a parallel implementation of the legislation and updates of the contracts between different entities (e.g. cloud service providers, customers, governments, private companies).
Comments
Leave a comment