While the ongoing AI arms race is focusing on accelerating the production of models to gain a lion share of an increasingly valuable market, are the right principles in place to govern and regulate the technology ethically?
At MWC Barcelona, four industry leaders discussed and analysed what regulations are already in place presiding over AI, what are some of the best business practices companies can apply, and why it may all boil down to common sense prevailing.
Claudia Nemat, Board Member at Deutsche Telekom, detailed how the introduction of the EU AI Act has made the continent a major player in the global market, but insists it must be able to bring forward new talent to become a true leader.
She revealed what are some of the growing pains challenging businesses when it comes to complying with the newly-introduced regulation, and how Europe can become a key player when it comes to fostering innovation with a regulated backbone.
Nemat said: “I believe Europe has to become a continent of chance again, hope and AI execution. We need more private and public capital, less documentation rules and less bureaucratic requirements.
“We need to see how innovation thrives in a vibrant ecosystem, an ecosystem that involves leading companies, startups, smaller companies and academia. In Europe, we should be proud of our leading companies and should stop making them smaller. For big companies, we need to become more courageous when it comes to working with startups.
“Europe should become the continent that attracts the best talent in the world with a welcome culture. My grandfather lived in the Nazi regime, and between 1933-1944 the best talent went to the US.
“My advice would be, for anyone who loves to develop startup technology, ethical AI, without getting overregulated, because this does not happen with rules, come to Europe and we promise to be less complicated and turn it into a narrative for the best talent in the world because we believe we solve problems for the betterment of mankind. We need to get back to this narrative.”
There is no trade off for Responsible AI
When the conversation turned to best practices for businesses to implement responsible, ethical and trustworthy AI, Richard Benjamins, CEO of RAIAGHT.ai, emphasised that this is not a trade off whilst simultaneously innovating.
He asserted that businesses should focus on both negative and positive impacts on their AI models and practices, focusing on designs that can become cheaper and more efficient when it comes to market launch.
Benjamins stated: “Ethical AI, trustworthy AI, responsible AI, for me they all combine within the same idea. It is not only about limited business protocols, it’s about looking at this from a wider context.
“Responsible AI, we usually speak about how to avoid negative impact, the ethics around being responsible. There is also an opportunity to create a positive impact. For me responsibility is not only avoiding negative impact, but you also have a responsibility to create a positive impact.
“This goes beyond business objectives, it involves society, climate, inclusion, so there are a lot of problems we need to solve.
“If you don’t have governance, you have no way to execute those business promises. Governance is about defining the process; you have to register your AI systems, you have to analyse the risks, you have to think about mitigating these risks or even prevent that risk.
“Another topic of discussion is also how do you be responsible while innovating? I think this is not a trade off. If you do responsible thinking by design, finding a risk earlier on, it is actually cheaper and faster to mitigate it rather than going as quick as possible to the market, then the risk happens and you go through reputation impact, and then apologising.
“It is not a trade off, this is something you have to do together.”
Why common sense may prevail
Whilst it is understandable from a company’s perspective that formulating an ethical, responsible AI framework can be challenging in a market that is rapidly developing, maybe it requires more common sense than they may think.
This was echoed by Francisco Jose Montalvo Abiol, CDO Director for Telefonica, who revealed that his company has been applying more ‘common sense’ when it comes to processing AI models and frameworks.
“One of the latest elements we have implemented is common sense and I know this may sound stupid, but most of the time when operators come to Europe, they want the simulations to happen in one minute, and this usually happens in one day,” said Montalvo.
Montalvo also gave some more clarification from his experience working with companies who may have apprehension when it comes to developing AI models. He assured the audience that it is okay to fail, but waiting for a perfect model will do little to progressing responsible AI.
He added: “I think we need to be a little less scared. Sometimes, it’s okay to fail, because only trying will help you develop the next model and for that you need to fail, but you need to be careful to not fail too much in terms of not creating a big hole in security.
“But we need to have a very honest conversation with our companies to say ‘this is not risk free, but still want to develop this. It’s not easy, of course, because it has a reputational impact. I think waiting for a completely safe model is something that is not going to work, and it is not going to make us progress. We need to progress.”
AI was the overwhelming key feature at MWC in Barcelona this week, and many companies will have converged at the event to gain a better understanding of forming their own responsible AI frameworks.
Whether they may be a big tech company like Google, or a startup just getting their feet wet, they all share the same apprehension when it comes to forming a secure framework that does not harm their reputation and protects their customers over a technology that can significantly shift within a day.
But as Montalvo noted, maybe this process needs a little more common sense and a confidence to fail before responsible AI can be crafted to be the industry standard for the betterment of businesses and customers alike.
