By Jeff Smith, Vice President of Systems Engineering APJ, Nutanix
Avoiding building bias into AI
Now that we’re in the data-building phase of AI, businesses need to be mindful of the biases people take on and avoid incorporating them into the AI algorithms that are fast becoming important.
We conducted research recently that showed Australian organisations were embracing AI with almost three quarters of respondents reporting a positive impact. Business and society are becoming increasingly comfortable with the technology, but there are still concerns and this is an area that needs to be monitored closely in 2019.
Imagine AI was at the heart of a screening system at an airport, or to put forward candidates for a job interview.
There is also a security aspect to this – even if developers can build the right algorithms that treat everyone fairly, there is a danger that the ever-increasingly sophisticated hacker will find a way to infiltrate systems and change algorithms to incorporate whatever biases they choose. The societal impact of this could be detrimental.
The flipside to this is that AI gives us the opportunity to remove our bias and understandably limited thinking from important decisions. We’re all guilty of making decisions based on a limited number of factors and relatively small sample and experience size, but AI has the power to look at things in black and white, taking into account any number of factors and data to determine the best outcome.
Think of an experienced surgeon looking at an X-ray to decide how he or she approaches the surgery. There are years of experience, training, understanding and more percolating to help the surgeon make this decision. But what if they could be assisted by AI that could analyse ALL similar surgeries conducted in the past 10 years? How valuable could this information be for the surgeon before they begin?
This is looking down the line further than 2019, but it’s important that organisations start thinking about these possibilities while we’re in build phase for future AI.
How A/NZ telcos can take on autonomous vehicles
Incidentally, telecoms companies across Australia and New Zealand have put themselves in a position to capitalise on innovative, data-intensive technologies such as driverless vehicles that will develop.
The large networks of exchanges and points of presence (PoPs) across each country all contain edge data centres, the kind of infrastructure that will be paramount to delivering these kinds of technologies. Why? Many of these burgeoning technologies will require a distributed network of nearby data centres to run. Speed and latency management are essential for these technologies to work.
Telcos’ existing network of data centres could easily be upgraded to accommodate this. With autonomous vehicles gaining ground in A/NZ and current airport trials in both countries likely to hit public roads this year, this opportunity could be very close.
Year of the multi-cloud
After much talk, 2019 should see the conversations businesses are having about cloud journeys turn to multi-cloud.
Businesses – and our wider societal mentality – are pushing against the idea of being locked into anything. We’re seeing the term ‘no lock-in contract’ appear much more as marketers seek to capture the modern customer.
This should be no different for businesses and the question of ‘which cloud is right for my organisation?’ is no longer the right question. It should be ‘which cloud should I use for the specific application I want to run?’. A better balance with hybrid cloud, which our research showed last year was the preference of more than 90 per cent of organisations in A/NZ and around the world, will lead to better business outcomes in 2019.
By Carlo Nizeti, Systems Engineering Director A/NZ, Nutanix
Say hello to your latest buzzword, AuT or ‘Autonomous Things’. After spending many years being hyped, Internet of Things is widely accepted – we know what it’s about even though we’re just getting started. Nonetheless, it’s time to make way for Autonomous Things as the new buzzword do jour.
Essentially, it refers small, autonomous robots that promise to seamlessly interact with our everyday lives. For example, some hotels are starting to use small robots that deliver toiletries and carry bags to rooms, spelling bad news (or perhaps, a more evolved role?) for the porters. Australia and New Zealand’s propensity for fast technology adoption could see it take off.
This is more of a gimmick example, but like all technological developments, businesses in A/NZ will start to mine it to help reach business goals such as more flexible, mobile workforces and increased use of data analytics, as well the more obvious: making the customer happier. Whatever happens, don’t be surprised to see AuT take off and gain some hype in 2019 – Gartner has already flagged it as its number one strategic technology trend to look out for in the year ahead.
A more sophisticated hacker
2018 was a huge year for cybersecurity in A/NZ, with the Notifiable Data Breaches scheme coming into effect in Australia, and the launch of New Zealand’s 2018 Cyber Security Strategy and Action Plan 2018. And let’s not forget Europe’s GDPR, which has a direct correlation to how companies operate in A/NZ.
This year is unlikely to see cybersecurity removed from the headlines. The sheer amount of data we’re creating is now enabling hackers to be far more knowledgeable and creative in their approach when targeting A/NZ businesses.
We’ve all seen the overtly obvious emails come through that promise riches for facilitating the transfer of money, but more sophisticated cybercriminals are now taking advantage of the data deluge to target businesses and staff referencing real customers, partners and information only people within the business should know.
This spells danger – if a hacker can imitate a colleague effectively, our spider senses won’t activate before clicking on a link or an attachment.
A report commissioned by Microsoft last year indicated the direct economic loss of cybersecurity incidents on Australian businesses is $29 billion per year. The costs of a breach – which can stretch well beyond financial damage to destroy relationships, trust and even the lives of those at fault – are simply too high and businesses and staff have an onus to up their security and awareness to protect against these attacks.
Soft landing for AI
Many of us still see artificial intelligence (AI) in terms of autonomous robots that can mimic humans, steal our jobs and even kill us – a likely result from the technology’s exposure in Hollywood over the years.
But the physical form of AI is still far down the track – connecting a sophisticated level of the technology to even simple physical movements such as picking something up is a gargantuan task, and hence why AI development in 2019 will continue to be centred on software.
Last year, we released our first annual Enterprise Cloud Index, which showed that despite widespread concern over AI, almost three quarters of mid-large organisations in A/NZ were benefiting from the technology. Only four per cent of organisations reported a negative impact. This is prime time for organisations to adopt a positive AI culture.
The area where I think we’ll see the most development in is AI-powered data analytics, and this is largely down to data. As we continue to create untold levels of data, we’ve made it impossible to manually sift through it to gain useful insights. Even ‘traditional’ analytics tools are becoming obsolete as, without a machine-learning component, they can’t keep up with ever-changing sources, types and volume of data.
Unsurprisingly, banking, finance and other industries with high volumes of data will likely be the first cabs off the rank in driving AI-powered analytics.