AI governance
To solve some of the common problems CSPs face when implementing AI projects , and t0 boost the chance of success , TM Forum members have developed multiple AI governance resources :
• AI Checklists are designed to provide a safe and effective framework for AI development at scale .
• The AI Canvas is a lightweight structured template to use at the beginning of the process to assess challenges and the appropriateness of applying AI to solve them . The canvas also helps to identify gaps and risks .
• AI Model Data Sheets help with documenting models and capturing information such as how they got built , what they ’ re for , their weaknesses and limitations , and how well they perform . Source : TM Forum , Counter intelligence : using AI to improve customer experience
[ Automated Intelligence Management ] can intuitively recommend and drive new dynamic services as new software containers without the traditional need for monolithic end-to-end firmware upgrades .”
“ Beyond the current capabilities of most Wi-Fi mesh networks , a more intelligent CPE , as envisioned by Gamgee , can offer a wide range of additional services and functions ,” adds Hendricks . “ These include IoT security with anomaly detection , device identification ( especially when MAC addresses are randomised ), home security systems , elderly care services with fall detection and activity tracking , and home automation utilising Wi-Fi sensing .”
“ We could see ‘ Tolkeinisation ’ – that is , one AI to rule them all ,” says Cottam . “ This will enable a broader range of services to be supported seamlessly and easily . Like Alexa or Siri , the AI might have their own personality in order to provide a friendlier interface . Unlike these smart assistants , she will be customisable and more proactive . The household AI could evolve to become a true companion that not only accurately recommends content to watch , but is also able to comment on or appraise films in the same way a human companion would .” TRANSFORMATIVE . “ The CPE market is undergoing a transformative phase , and the prpl Foundation has been pivotal in driving it ,” claims Bensaid . “ We are transitioning from a fragmented market where innovation was costly and disparate , to a market guided by a standardised approach . This shift will enable the emergence of new services like a more automated home environment , enhanced security alongside trustworthy and ambient connectivity . As AI and connected technologies continue to advance , the potential of intelligent CPE is vast and promises a more sophisticated and user-friendly home networking experience . The contributions of prpl are underpinning this revolution .”
In addition to his prpl role , Bensaid , in his day-to-day job as SoftAtHome ’ s VP Marketing , notes that nowadays , CPE are very generic and are all equal . “ In the future , we can imagine a personalised CPE that automatically adapts to the user and their context . Why not adjust the configuration automatically using AI ? You are a gamer ; your CPE becomes your best partner optimising the latency so that you can win . You are a binge watcher ; your CPE increases the bandwidth . You are working from Home , no problem ; here is the best configuration . AI would be an assistant that will automatically , and from behind the scenes , adapt your environment to your usage .”
“ There is an endless list of services that an intelligent CPE can provide in the home ,” declares Ferré . “ Just as set-top boxes have long been the platform for operating all video services , CPE and intelligent extenders could become the best platform for operating a fully connected smart home .” TRUST . There may already be a trust issue with AI and the consumer . If so , how transparent must operators be about its use , and what are the safeguards ? “ There is a trust issue with AI being seen as a ‘ Big Brother ’ in an Orwellian context , as it can recognise and perform actions automatically using semiintelligence on behalf of the subscriber ,” states Walls . “ Whereas adopting a standardised , containerised services approach would bring the desired fluidity and control that service providers crave . A key message is that AI has significant benefits to subscribers ’ Quality of Experience at a network service level but not at a personal data level .”
“ Convenience vs Trust is one of the scales that may tip towards convenience with younger demographics vs older ones ,” suggests Cheevers . “ When you have less to lose ( like money or privacy issues ) then you tend to swing more towards convenience .
“ AI must simplify things and help give context and options , but not ‘ run the show ’ or the home .”
- Jean Ferré , Vantiva
When you have more to lose , then you add more safeguards and trust elements . The convenience and improvement factor has to outweigh the trust issue from the consumer . But in all cases , the consumer has to be aware of the access they are allowing to get this benefit .”
“ Although the industry hypes AI , with ChatGPT and similar technologies gaining attention , the general audience is often unaware of AI ’ s current and potential role in their lives ,” notes Hendricks . “ People generally lack awareness until something goes wrong , as seen with cybersecurity . Legislators recognise the risks associated with AI and are working on passing relevant laws in the near future . However , legislation often struggles to keep up with rapidly evolving technologies . Companies utilising AI in consumer services have a responsibility to provide transparency and ethical use , even before legislation is in place . Operators must be transparent about AI ’ s application , its benefits , and potential risks .” CHANGE . “ With every technology change there is always nervousness , fear and excitement in equal measure amongst customers ,” observes Cottam . “ People are not unwilling to try AI – whether via their smart speaker or via sites such as character AI – but it ’ s essential that operators are clear about how AI will benefit customers , not over-promise on the technology , and reassure customers that their data is not at risk . Although we ’ re still at the start of this journey , TM Forum members are already collaborating to develop and share best practice , and have developed a range of resources to help operators adopt AI safely .”
“ AI is everywhere , and we don ’ t really know ,” says SoftAtHome ’ s Bensaid . “ Operators are trusted partners that can help consumers navigate the digital world confidently by securing the connections and ensuring that nobody uses AI or any other technology to steal data or influence behaviour . More than being transparent , Operators should step up and become customers ’ partner of choice to secure their digital lives .”
“ Our home is the most private and important place in our life ,” asserts Ferré . “ A home is a place where we should feel safe , secure , and protected ,” he says , adding that “ the home should give us two guarantees : To respect our privacy and use our data as context to give us the best possible insights and services , but not to trigger tracking or advertising , and to support a better home experience without subordinating humans to AI .”
“ AI must simplify things and help give context and options , but not ‘ run the show ’ or the home . With the help of CPE manufacturers , network service providers can provide a helpful solution in the context of trust between the home users and the NSPs ,” he concludes .
EUROMEDIA 15