AI as co-creator: building better software and better businesses
As we see uses of artificial intelligence (AI) proliferate in the economy, certain patterns become clear. The industries and firms whose digital transformation seems to be benefiting most are those using AI to make processes more efficient. Indeed, according to Deloitte’s State of AI in the Enterprise report, the most experienced AI practitioners rank making processes more efficient as their number one priority, closely followed by using AI to create new products and services.
To be clear, we are not quite at the precipice of artificial general intelligence, as featured in films like Her and Ex Machina. The reality is much more prosaic. Jeff Bezos says that AI’s primary contribution to Amazon so far is “quietly but meaningfully improving core operations.” To understand how businesses are successfully adopting AI, and how others can learn from them, let’s look at how AI is already contributing to improved processes and the creation of better digital products.
One of the four key areas of AI — alongside machine learning (ML), deep learning, and natural language processing (NLP) — is computer vision, which describes the ability to extract meaning and intent from visual elements, such as text, images, objects, interfaces, etc. While Tesla’s autopilot feature is probably the best-known application of computer vision, it’s possible that software that you use every day has benefited from computer vision in some manner.
Finding issues before users do
For example, computer vision can be used to interpret and interact with software, just like humans do. In AI parlance, this means using intelligent agents to explore a software environment and then determine actions to be taken. Why would we want to do this? Well, as software becomes further entrenched into our daily lives, we need to make sure that not only does it function as intended, but that it is usable and accessible by a range of real people on various devices. By using intelligent agents to test software and identify issues that impact the user experience, developers can improve software by leaps and bounds.
Enabling AI capabilities through the cloud
If we broaden the scope from improving software to improving the adoption of AI more widely, the commoditisation of AI capabilities through cloud computing has created an interesting area of opportunity. The cloud is an enabling technology and can be the ideal vehicle for doing AI. As Tom Davenport, Fellow of the MIT Initiative on the Digital Economy, puts it, “The cloud is an amazing distribution mechanism for algorithms.”
It’s also vital to widen access to these cloud computing capabilities. Sebastien Krier, an AI policy expert at the Stanford Cyber Policy Center, suggests creating a pool of cloud compute credits for the UK’s R&D ecosystem. This pool of subsidised cloud credits “would allow researchers to work on socially beneficial AI applications, scrutinise commercial models, and undertake long-term research that the private sector would not naturally be incentivised to fund.”
To ensure that all sectors benefit from the advances of AI, it’s important to both widen access to critical AI capabilities in the cloud, as well as expand our idea of how AI can be used to build better software and more efficient organisations. Only then will we reap the rewards of this transformational technology.
Jaspar Casey, Product Marketing Manager at Eggplant
Katherine joined techUK in May 2018 and currently leads the Data Analytics, AI and Digital ID programme.
Prior to techUK, Katherine worked as a Policy Advisor at the Government Digital Service (GDS) supporting the digital transformation of UK Government.
Whilst working at the Association of Medical Research Charities (AMRC) Katherine led AMRC’s policy work on patient data, consent and opt-out.
Katherine has a BSc degree in Biology from the University of Nottingham.
- [email protected]
- 020 7331 2019