6G and the need for one single AI standard - can you believe we are talking about 6G?!
In November Industry Analyst and RMA Managing Director, Guy Redmill, moderated several sessions at this year’s Telco AI World Summit - and presented his thoughts on “Bringing AI into the standardisation fold with 6G”.
You can watch the video of the session: a mere seven minutes. And/or you can also carry on reading for a quick summary…
6G‘s Wikipedia page says that 6G broadband cellular networks will offer lower latency, massive amounts of bandwidth and speeds significantly faster than 5G - around 95Gb/s.
With commercial availability of 6G projected for the 2030s, research is full steam ahead within the vendor and academic communities.
There’s speculation that 6G won’t rely on wired networks at all with devices acting as antennas using networks not under the control of a single MNO. If true, then of course, the marketing of this technology to the general public will need to be carefully planned as we’ve all seen the resistance to 5G by vocal campaign groups.
Meanwhile the ITU is working on recommendations for 6G, the IMT 2030 (International Mobile Telecommunications). To be clear, the IMT simply sets out desired performance capabilities. It does not specify how this should be achieved or provide a standard.
With 5G the industry avoided the creation of different approaches by following a single implementation of the IMT recommendations – unlike previous generations of mobile technology (4G over WiMAX anyone?). While this brought benefits, our recent research has led us to believe that 6G probably won’t follow a single standards path with Governments (many of whom are taking a keen interest though not as uninformed as Trump’s assessment of 6G…), industry verticals and new groups creating their own standards.
Prediction: RMA expects 3GPP2 to be revived, at the very least
Interesting stuff which will be covered - and more - in a new book on 6G that Guy has contributed to which is in the publishing pipeline.
AI in 6G will need standardisation no matter what
Back in 2017, Deutsche Telecom’s senior executives told industry news and analysis site, Light Reading, that the company could only succeed through a process of "brutal automationrutal automation" with networks that can function with "no human involvement".
Full automation is just one of the goals for 6G networks - and crucial for the delivery and management of both is AI to enable the processing of increasing volumes of data as well as service problem management, policy-based management and orchestration.
But here’s the challenge for the ITU, IEEE, ISO, and Governments: the techniques for delivering AI in 6G don’t exist yet.
But that doesn’t mean there aren’t already all sorts of concerns regarding the application of AI, how it processes data – and how it should be aligned with other data processing regulations.
However, efforts to standardise AI - as it might be in the future - have started with the focus firmly on security, trustworthiness and the use of data.
A tricky task as we currently have an open, innovative, unregulated ecosystem building and deploying AI-based systems.
But a task that none the less is as vital as it is complicated for our future 6G networks.