The Big AI Con: Are You Really AI?
Originally posted on N2 Technology on 03/03/2021.
Artificial intelligence has infiltrated many aspects of our lives. From Siri and facial recognition in our smartphones to virtual healthcare and the fight against Covid-19, AI Technology has rapidly changed the way we live and work, every day. The AI market is forecast to reach $15,7 trillion by 2030, so it’s no surprise startups are determined to advertise their Artificial intelligence capabilities - regardless of their extent, or definition.
The biggest complication when it comes to AI is its definition. It depends on who you ask. Back in the 1950s, Minsky and McCarthy described artificial intelligence as “any task performed by a machine that would have previously been considered to require human intelligence,” however, modern interpretations are more specific. According to the European Parliament, AI exudes the “ability of a machine to display human-like capabilities such as reasoning, learning, planning and creativity.” But where does automation end, and Artificial intelligence begin?
AI can perceive its environment, cultivate conclusions, solve problems and act to achieve set goals. These systems are able to adapt their behaviour by analysing the effects of previous actions and work autonomously. Automation, however, is more simple. It focuses solely on streamlining repetitive, instructive tasks with the purpose of saving time and money spent on monotonous, voluminous tasks. In turn, allowing employees to apply themselves to more complex processes.
But here’s the thing: AI is a buzzword that’s exciting investors - and startups have caught on.
The buzz of AI
New technology usually invokes great excitement. This is because tech can be truly transformational, and this time, AI is no exception. Increasingly gaining popularity across all sectors and quickly becoming the preferred tool for earning a competitive edge, AI offers efficiency, quality, speed and more:
reduces employee work pressure,
offers maintenance of inventory,
easy analysis of huge databases,
innovative decision making,
help in market analysis,
assurance of security maintenance
These capabilities are transferable across many industries. In customer experience, AI utilises customer data to outline customer journey maps, reduce wait times, simplify transactions and enhance customer engagement. In healthcare, AI can analyse big patient data sets to deliver better, faster and cheaper healthcare - as well as support preventative medicine and new drug discovery. IBM’s Watson can pinpoint treatments for cancer patients and Google Cloud’s Healthcare app can help health organisations collect, store and access data. In finance, AI can better predict and assess loan risks, help improve load underwriting and reduce financial risk.
With so many transformative benefits coming from AI, it’s no wonder investors are interested. However, the question begged by Forbes; ‘is venture capital investment in AI excessive?’ stands appropriate. Sure, Artificial intelligence is - at times - breaking unfathomable boundaries, but it’s not new. AI has been around since the dawn of computing, but both startups and VC funds are cashing in on the hype; “startups are playing into this trend and raising more money than ever, as long as they have AI or cognitive technologies in their business plans or marketing material,” says Kathleen Walch. “[And] VC funds themselves are raising skyrocketing levels of new capital if they focus their portfolios on AI Technology and related areas.”
AI companies raise more money across fewer rounds
According to BuyShares.co.nz, AI startups raised $9.9 billion in the second half of 2020 - 15% more compared to the same period in the previous year. Despite the crisis of Covid-19 causing a sudden interruption between January and June 2020, investment in AI funding sustained momentum. In the third quarter alone, $69.6 billion was raised in cumulative funding - a $4.7 billion increase in three months. $5.2 billion was raised between September and December 2020 - 40% more than the same period a year ago, with a cumulative value of $74.8 billion cumulative funding.
In 2019, London-based VC fund, MMC, found that out of 2,830 startups in Europe classified as being AI companies, only 1,580 accurately fit that description. In many cases, the label which refers to computer systems that can perform tasks normally requiring human intelligence, was wrong.
“We looked at every company, their materials, their product, the website and product documents,” says David Kelnar, head of research for MMC. “In 40% of cases, there was no mention of evidence of AI. In such cases, companies that people assume and think are AI companies, are probably not.” But here’s the thing, these startups are not necessarily promoting themselves as AI firms - they’re actually being classified by third-party analytics websites, and Kelner believes that in most cases, these startups are aware of their classification. They just fail to correct the listing because there is no incentive to.
Companies labelled as being in the field of AI generate extra press, and attract 15% to 50% more in their funding rounds than other technology startups. Therefore, if a company fails to fall under the AI umbrella, there’s potentially less investment down the line. Ultimately: it pays to brand your business as AI. For instance, specialist engineers are a lucrative role, so they require higher salaries. But more importantly, Kelnar resolves, these figures are a “reflection of the dynamics of supply and demand.”
There’s little impetus for correction. Why would a company risk losing its AI title when even the most technical CEOs struggle to fully comprehend artificial intelligence? Princeton professor and AI ethics commentator Arvind Narayanan suggests in his findings that some companies exploit this public confusion by slapping an “AI” label on their products — a term he describes as “AI snake oil” - or otherwise recognised as AI-washing, or Fake AI.
Watch out for AI Washing or Fake AI
In a growing trend, it appears that software companies are “exploiting the current artificial intelligence craze by exaggerating the scope and capabilities of AI in their products.” Coined by Gartner using their Hype Cycle tool that measures the growth and decline of products as they mature, “AI Washing” describes the process of ‘overhyping’ - similar to the term “greenwashing”. Gartner found that more than 1,000 vendors say their products employ AI, but many are “applying the AI label a little too indiscriminately.”
“Human-assisted technology is not artificial intelligence”, says Lauren Hamer. Let’s take a look at three ways companies loosely define AI for their benefit:
Voice recognition software: a Chinese company boasted fearless Artificial intelligence claims in 2018. During an audience presentation, the company demonstrated how its automated voice software could conduct real-time language interpretations from English into Chinese. After scoring millions in funding to create next-generation smart speech technology, a previous employee claimed to be one of the human ‘interpreters’ simultaneously transcribing speeches behind the screen.
App development outsourcing to tech engineers: an Indian-based app development platform won nearly $30 million in funding after claiming that users can build up to 80 percent of a mobile app - from scratch - in about an hour. However, the company actually employed human engineers in India to assemble the code - rather than using Artificial intelligence features.
AI personal assistant software using humans: have you ever thought that your automated assistant sounds just like a real person? That might just be because they are. Bloomberg discovered that mild-mannered helpers designed to schedule meetings or order food aren’t always entirely robotic. Former employees have revealed that there is a human behind most automated actions. In one company, a human was paid to highlight phrases entering in the system to help the computer ‘learn’ and formulate proper responses - which led to disgruntled employees from a competitor calling the tech “smoke and mirrors”.
It’s argued that humans are needed in the initial stages to fine-tune algorithms in the backend. Without the malicious intent to deceit, companies try to ‘fake it until they make it’ before they run out of money. The risk of removing the Artificial intelligence label can hinder the buzz and sales. As the Guardian put it; “it's cheaper and easier to get humans to behave like robots than it is to get machines to behave like humans.”
But does this benefit investors? It’s believed that people tend to disclose more when they think they’re talking to a machine, not a person - and any marketer knows the value of information - especially when it's about a customer.
Conclusion
Artificial intelligence is one of the most misused terms in tech today - and all parties are complicit. “Zealous marketing departments, capital-hungry startup founders and overeager reporters are casting the futuristic sheen of AI over many products that are actually driven by simple statistics - or hidden people,” explains Kaveh Waddell.
But does it really matter?
The problem lies in “overinflated expectations for technology. It undermines public trust and potentially sets up for backlash,” continues Waddell. This leaves customers distrustful of products and practising copious levels of caveat emptor (let the buyer beware). The investor overpays, and the vendors run a catastrophic risk to their reputation. If the exploitation of the AI buzzword continues, the consequences could damage the entire market, curtailing innovation and technological advancements. As the old saying goes: ‘once bitten twice shy’ is relevant to deceived investors.
It’s imperative that buzzwords don’t cloud capabilities. AI comes with great expectations, but not all AI systems and services are created equally. Sometimes basic automation - or a human - might just be more effective than AI, so why not advertise accordingly? The risk of deceit, in the long term, outweighs the reward.