The entrepreneurial landscape has largely benefited from on-demand cloud computing services, provided by Amazon Web Services (AWS) and others. The accessible and instantaneous supply of computing power means that companies are able to focus on what they do best. They are spared from costly infrastructure investments and inflexible long-term planning. AWS has significantly lowered the barrier to entry for tech startups and marketing tech vendors. Companies can set up shop and then scale their operations as needed, with a relatively high degree of security and reliability. AI might have a similar effect on businesses.
Some tech leaders have forecasted that the best AI will soon be accessible to all, for a reasonable price, in the style of cloud computing services. If this comes to pass, more companies will be able to extract marketing insights from their data.
At the present time, some proprietary versions of AI are only available with enterprise software subscriptions. These systems can assist with correspondence, sales cycles, media planning, competitive intelligence, business predictions, and other efforts. You could argue that these forms of AI are gunning for certain people’s jobs. But oftentimes, there’s a significant, enterprise-level price tag attached. That price tag is staving off technological unemployment for some workers. But it’s also restricting the business and economic benefits of AI.
Expect those numbers to drop. Expect the early adopters to give way to a majority. This is technology history 101 and the same principles apply here.
Some AI breakthroughs are emerging through open frameworks that emphasize ethics and collaboration. Developers in this camp are cognizant of their work’s potential impact. They believe that the need to generate financial returns could ultimately constrain the development, ethics, and distributed benefits of AI. Others want to make a profit, by addressing customer pain points. And others want to build a weapon, equating the field to a new global arms race. It isn’t yet clear how these differing philosophies and uses will intersect. There are developers who see all sides of the situation. There are even developers who see a bad outcome as a near certainty. Yet their work goes on.
Provided apocalyptic scenarios are averted, it’s possible that AI will become a widely accessible business enhancement, like AWS. Companies will have their own datasets and features but they will all be plugging their numbers into a nondiscriminating AI system, provided by a tech giant like Amazon or Google or IBM.
The precedent is clear. The AWS website explicitly describes the economic effects of its services. One webpage states, “AWS continues to lower the cost of cloud computing for its customers. In 2014, AWS has reduced the cost of compute by an average of 30 percent, storage by an average of 51 percent and relational databases by an average of 28 percent.”
To back up its claims, AWS provides a calculator so that tech companies can easily compare the cost of running their applications on-premises or in a colocation environment against the price tag of AWS.
It’s easy to see the same thing happening with AI on a large scale. It’s already going in that direction. Soon, even small businesses will be using a calculator to compare the cost of an in-house AI against the Amazon-provided one.
According to recent analysis from Deloitte Global, companies will accelerate their usage of cloud-based AI in 2019. The cloud availability will drive more full-scale AI implementations, better ROI, and higher AI spending.
But a skills gap might prevent organizations from integrating AI into their processes and workflows in the most effective way. AI is also laden with risks, even beyond the existential ones. Sloppy algorithms and biased or insufficient datasets could harm corporate reputations. Flawed data practices might also result in legal liabilities.
AI remains a risky proposition. Over the long term, AI will allow businesses to operate more effectively. They’ll save money. But society might have to incur significant damages upfront, in terms of ethics and employment, before the technology’s benefits can be fully realized and distributed.
