Given the disruption that has persisted throughout 2020, we commend AI Summit for going ahead virtually this year. Our thoughts go out to the organizers who have clearly put in significant effort to create an environment that allows the best and most inspired minds in AI to connect with the most curious.
Brainpool was fortunate enough to attend every session in the Summit’s first episode: Global Game Changers, Delivering on the Promise of AI. The episode focused on three overarching themes: Equality, Governance and the Ethics of AI. Together these offered an insightful overview of how we should appraise and approach the implementation of Artificial Intelligence in business, society and our lives.
Below is an overview of the topics covered and our thoughts on how they might impact businesses seeking to develop AI.
Equality in AI:
With a range of sessions about how AI might impact or be impacted by human biases, it is becoming increasingly clear how we should take action to ensure AI works for everyone, regardless of culture, gender or ethnicity.
Rediet Abebe, Harvard Fellow and Co-Founder of Mechanism Design for Social Good, deserves special mention as not only the first speaker of the Summit, but as someone whose research shows the practical methods we might use to route out inequality. After this episode we reviewed her paper on the ‘Roles of Computing in Social Change’ to get more information. Her recommendation of using technology as a diagnostic tool for understanding inequality is one that all companies should consider. If you have the ability to build these systems, it can become an essential tool for developing a fair and equal modern business. Organisations have a responsibility to adopt these tools, with equality in mind, to ensure that bias and discimination is not reinforced, compounded or further institutionalised.
Governance of AI:
The role that governments should play in the regulation of AI was also a hot topic. The need to strike a balance between ensuring that innovation isn’t stifled yet effective regulation is in place is crucial. What that balance looks like is up for debate.
Moderated by Meriem Menani, Public Policy Consultant The Economist Group, the session on technology governance provided insight into the importance of multi-stakeholderism. Francesca Rossi, IBM fellow and AI Ethics Global Leader, highlighted the fact that emphasis should be placed on regulating high-risk applications of AI, rather than AI itself. As such, Elissa Stromme pointed out that since applications of AI present themselves in already highly-regulated areas such as finance, it is vital that existing regulation is constantly updated to ensure effective governance.
Posed with a question of whether AI should be used for the good of society, Damian Boesalager MEP suggested that we should ensure technologies are developed with wider benefits in mind from the out-set - against this aim, governments are well-placed to oversee development. Importantly, the guiding principles that shape this regulation must take into account a variety of voices - a multistakeholder approach is vital. This approach must also ensure that internal teams are responsible for operationalising principles, not just outlining them, as was stressed by Francesca Rossi. In doing so, these teams will develop an understanding of how AI-based applications, governed by guiding principles, can be easily integrated into existing processes.
Ethics of AI:
The ethics of AI continues to dominate thought leadership within AI. Sessions in episode one focused on how building ethical Artificial Intelligence can bolster public trust of the technology. Professor Teemu Roos was keen to point out how fear of AI will have a negative impact on its adoption and wider innovation. Democratising education and knowledge of AI is clearly essential within this and should not remain a privilege of the few.
Ajay Bhalla, President of Cyber and Intelligence Solutions at Mastercard, had perhaps the clearest philosophy of how to develop trustworthy AI. His key principles of ethical, fair and secure systems are concepts that should be considered by every company seeking to implement the technology. As always the challenge remains that ethics depends on your business, culture, or outlook.
Discussions about ethics are often interesting and engaging but often there is no clear consensus on what actionable steps should be taken. Perhaps that’s because AI is for all intents and purposes in its infancy. The capability of the technology is in its early days and as with most innovation, the ethics and regulatory considerations lag behind overtaken by the far more stimulating pursuit of progress. But whilst no united ethical or legal framework looms on the horizon anytime soon, we’re delighted to see that discussion and thought leadership hasn’t slipped into the background during a time when the struggles of humankind could have motivated many to want to play fast and loose with ethics and morality with the intent to make the world a better place.
We look forward to subsequent conversations that focus more on the practical and real-world considerations of AI.
Brainpool will be attending the next episode of the summit on the state of AI. We are excited for the University of Toulouse ANITI Chair, César A. Hidalgo’s talk on how humans judge machines, to provide us with more insight into what we consider effective and ethical Artificial Intelligence. Professor Debjyoti Paul’s talk on finding needles in the haystack of unstructured data also presents itself as a fantastic overview of the business value to be found in properly cleaning, formatting and structuring data.
Remember that you can catch up with us in our cyberbooth for more information on Brainpool, to discuss all things AI and how we Brainpool can help your business solve its data, technology and AI challenges.
We will be back every week with our top takeaways from each episode.
Written by Anjali Kapila, Dominic Richmond and Clayton Black