Digging in the State of AI at Innovfest Unbound
Appier’s Charles Ng speaks about the state of AI at Innovfest Unbound
Yesterday, Charles Ng, Appier’s VP of Enterprise AI, joined a panel of AI experts to discuss the current state of AI at the annual Innovfest Unbound conference in Singapore. In front of a full house, Charles and his fellow panellists, along with moderator Tim Culpan of Bloomberg, spoke on topics including challenges of AI implementation, mistakes in applying AI and novel ways we might see it being used in the future.
Bin Ru Tan, CEO of OneConnect, which applies AI in finance, shared that data remains a challenge, particularly when it comes to moving data across borders for everyone’s benefit, due to differing rules and regulations. However, Ng emphasized the significant role of data in effective AI implementation, and encouraged the audience to take a broader approach to collecting and leveraging data.
It is well known that data is a cornerstone of AI, and there is a common phrase of ‘garbage in, garbage out’, meaning that your AI outputs are only going to be as good as what you put in. The truth, according to Ng, is that data is not as black and white as this. Most people trying to leverage data won’t have ‘garbage’ data, rather they will probably have ‘imperfect’ data. The right approach is ‘better in, better out’. He said that it is a mistake to wait until your data is ‘good enough’ – there will always be more data, so it is a never-ending story. Start now, and work to make your data more robust and easier to manage over time.
When speaking about AI, it is inevitable that a discussion comes up about whether (or when!) AI will overtake humans. While this idea has become popular in films, for example, all the panellists agreed that we are quite some way from this!
Luis Gonzalez, Principal Director APAC at Element AI, and Tonio Buonassisi, MIT Professor & Principal Investigator at SMART SG, agreed with Ng and Tan that we will continue to see AI systems that operate exceptionally well in specific domains, helping people do their jobs better, instead of doing their jobs for them. The best AI will be designed with the human end-user in mind, and it will work alongside human intelligence for optimal outcomes.
The panellists wrapped up by each sharing what they expect in terms of unusual or surprising ways we might see AI applied in the future. Tan is optimistic that we will see more cross-border benefits when it comes to sharing data while Buonassisi is excited to see much more customization of products and services for individuals. Gonzalez shares the same view and sees a lot of opportunity in personalization; and Appier’s Ng expects to see more interesting applications of AI in generating new things such as images, stories, etc., going beyond AI’s current core strength of prediction and detection.
Thanks to all who attended our session, and look forward to seeing you at the next event!
* Read our blog for more insights on AI and how it can help with your business!